perf(plugins): dramatically speed up plugin manager tab load time (#333)

* fix(cache): check odds keys before generic live check in get_data_type_from_key

Cache keys like odds_espn_basketball_nba_<id>_live contain both 'odds'
and 'live'. The previous ordering matched the generic 'live' check first,
returning 'sports_live' (30 s TTL) instead of the correct 'odds_live'
(120 s TTL). This caused the ESPN odds API to be hit every 30 s per live
game, frequently triggering the 3-second per-request timeout and returning
no odds data.

Moving the 'odds' check above the generic 'live' block restores the
correct 120-second cache TTL for in-progress game odds.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

* fix(display): use single-quoted HTML attributes for JSON hidden inputs

Placing |tojson output (which contains double quotes) inside a
double-quoted HTML attribute broke the attribute — browsers closed
the attribute at the first inner quote, leaving JS with an empty or
truncated value. JSON.parse then failed silently, leaving excluded=[]
so all Vegas scroll plugins appeared checked (included) regardless of
the actual excluded_plugins config.

Switch to single-quoted HTML attributes so the JSON double quotes
are valid inside the attribute value.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

* perf(plugins): dramatically speed up plugin manager tab load time

## Problem

The Plugins tab loaded slowly and inconsistently (5–30s depending on
cache state), with a blank spinner for the entire wait. Three root
causes:

1. **N+1 subprocess per installed plugin** — `_get_local_git_info` ran
   4 separate git subprocesses per plugin (rev-parse HEAD, abbrev-ref,
   config --get remote.origin.url, log --format=%cI). With 15 plugins
   that's 60 blocking subprocess spawns before the endpoint returned.

2. **Serial per-plugin loop** — the `/plugins/installed` endpoint
   processed each plugin sequentially: manifest read → git info →
   instance lookup → Vegas mode query, one plugin at a time.

3. **Serial JS loading** — the store search only started after installed
   plugins fully completed, so users waited for both round-trips back
   to back. No UI feedback during the wait.

## Changes

### Backend — src/plugin_system/store_manager.py
- Consolidate 4 git subprocesses → 1: branch read from `.git/HEAD`
  (file I/O, no subprocess), remote URL parsed from `.git/config`
  (file I/O, no subprocess), SHA + commit date fetched together in a
  single `git log -1 --format=%H%n%cI` call
- Existing signature-based cache already eliminates all subprocesses on
  warm hits; this change cuts cold-cache cost from 4 → 1 per plugin

### Backend — web_interface/blueprints/api_v3.py
- Wrap per-plugin work in a `_build_plugin_entry()` helper and execute
  it across a `ThreadPoolExecutor(max_workers=8)` so all plugins are
  processed in parallel instead of sequentially
- Fix double `get_plugin()` call per plugin (was called once for the
  enabled fallback and again for Vegas mode — now one shared call)

### Frontend — web_interface/static/v3/plugins_manager.js
- Fire `searchPluginStore()` and `loadInstalledPlugins()` simultaneously
  instead of waiting for installed to complete before starting the store
- After installed data arrives, call `applyStoreFiltersAndSort(true)` to
  refresh install/update/reinstall badges from already-cached store data
  (instant, no extra network call)

### Frontend — web_interface/templates/v3/partials/plugins.html
- Add responsive skeleton cards to the installed plugins section that
  match real card proportions (removed automatically when data renders)
- Replace the 5 featureless gray boxes in the store skeleton with 10
  structured skeleton cards matching the real card layout

## Measured improvement on Pi 4 (11 installed plugins, ledpi-ticker)

| Scenario | Before | After |
|---|---|---|
| Cold cache (first open) | ~8–15s | **0.9s** |
| Warm cache (git cache hit) | ~1–2s | **55ms** |
| UI feedback during load | blank spinner | skeleton cards |
| Store waits for installed | yes (serial) | no (parallel) |

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

* fix(plugins): harden git metadata parsing and plugin entry building

store_manager.py:
- Detect worktree/submodule .git files (gitdir: <path>) and resolve
  to the actual git directory before reading HEAD or config
- Wrap HEAD read_text in try/except OSError/NotADirectoryError so
  atypical repos return None instead of propagating exceptions
- Guard config url line split with '=' presence check to avoid
  IndexError on malformed lines

api_v3.py:
- Wrap _build_plugin_entry body in a try/except via a thin outer
  wrapper so a single plugin's failure doesn't 500 the whole endpoint;
  failed entries return None and are filtered by the existing [r for r
  in results if r is not None] step
- Narrow manifest except clause to FileNotFoundError, PermissionError,
  json.JSONDecodeError instead of bare Exception
- Validate manifest is a dict before calling plugin_info.update() and
  log a debug message when it isn't

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

---------

Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>

README.md

@@ -1,4 +1,5 @@
 # LEDMatrix
+[![Codacy Badge](https://app.codacy.com/project/badge/Grade/77fc9b446a5948e5b0aed7a7aaeb1bab)](https://app.codacy.com/gh/ChuckBuilds/LEDMatrix/dashboard?utm_source=gh&utm_medium=referral&utm_content=&utm_campaign=Badge_grade)
 ## Welcome to LEDMatrix! 
 Welcome to the LEDMatrix Project! This open-source project enables you to run an information-rich display on a Raspberry Pi connected to an LED RGB Matrix panel. Whether you want to see your calendar, weather forecasts, sports scores, stock prices, or any other information at a glance, LEDMatrix brings it all together.
 

config/config.template.json

@@ -126,6 +126,11 @@
             "buffer_ahead": 2
         }
     },
+    "sync": {
+        "role": "standalone",
+        "port": 5765,
+        "follower_position": "left"
+    },
     "plugin_system": {
         "plugins_directory": "plugin-repos",
         "auto_discover": true,

first_time_install.sh

@@ -1086,6 +1086,7 @@ SYSTEMCTL_PATH=$(which systemctl)
 REBOOT_PATH=$(which reboot)
 POWEROFF_PATH=$(which poweroff)
 BASH_PATH=$(which bash)
+JOURNALCTL_PATH=$(which journalctl 2>/dev/null || true)
 
 # Create sudoers content
 cat > /tmp/ledmatrix_web_sudoers << EOF
@@ -1101,10 +1102,23 @@ $ACTUAL_USER ALL=(ALL) NOPASSWD: $SYSTEMCTL_PATH restart ledmatrix.service
 $ACTUAL_USER ALL=(ALL) NOPASSWD: $SYSTEMCTL_PATH enable ledmatrix.service
 $ACTUAL_USER ALL=(ALL) NOPASSWD: $SYSTEMCTL_PATH disable ledmatrix.service
 $ACTUAL_USER ALL=(ALL) NOPASSWD: $SYSTEMCTL_PATH status ledmatrix.service
+$ACTUAL_USER ALL=(ALL) NOPASSWD: $SYSTEMCTL_PATH is-active ledmatrix
+$ACTUAL_USER ALL=(ALL) NOPASSWD: $SYSTEMCTL_PATH is-active ledmatrix.service
+$ACTUAL_USER ALL=(ALL) NOPASSWD: $SYSTEMCTL_PATH start ledmatrix-web.service
+$ACTUAL_USER ALL=(ALL) NOPASSWD: $SYSTEMCTL_PATH stop ledmatrix-web.service
+$ACTUAL_USER ALL=(ALL) NOPASSWD: $SYSTEMCTL_PATH restart ledmatrix-web.service
 $ACTUAL_USER ALL=(ALL) NOPASSWD: $PYTHON_PATH $PROJECT_ROOT_DIR/display_controller.py
 $ACTUAL_USER ALL=(ALL) NOPASSWD: $BASH_PATH $PROJECT_ROOT_DIR/start_display.sh
 $ACTUAL_USER ALL=(ALL) NOPASSWD: $BASH_PATH $PROJECT_ROOT_DIR/stop_display.sh
+$ACTUAL_USER ALL=(ALL) NOPASSWD: $BASH_PATH $PROJECT_ROOT_DIR/scripts/fix_perms/safe_plugin_rm.sh *
 EOF
+if [ -n "$JOURNALCTL_PATH" ]; then
+    cat >> /tmp/ledmatrix_web_sudoers << EOF
+$ACTUAL_USER ALL=(ALL) NOPASSWD: $JOURNALCTL_PATH -u ledmatrix.service *
+$ACTUAL_USER ALL=(ALL) NOPASSWD: $JOURNALCTL_PATH -u ledmatrix *
+$ACTUAL_USER ALL=(ALL) NOPASSWD: $JOURNALCTL_PATH -t ledmatrix *
+EOF
+fi
 
 if [ -f "$SUDOERS_FILE" ] && cmp -s /tmp/ledmatrix_web_sudoers "$SUDOERS_FILE"; then
     echo "Sudoers configuration already up to date"

plugin-repos/march-madness/requirements.txt

@@ -1,4 +1,4 @@
 requests>=2.28.0
-Pillow>=9.1.0
+Pillow>=12.2.0
 pytz>=2022.1
 numpy>=1.24.0

scripts/download_pixlet.sh

@@ -118,7 +118,7 @@ total_count=${#ARCHITECTURES[@]}
 
 for arch in "${!ARCHITECTURES[@]}"; do
     if download_binary "$arch" "${ARCHITECTURES[$arch]}"; then
-        ((success_count++))
+        success_count=$((success_count + 1))
     fi
 done
 

scripts/install/configure_web_sudo.sh

@@ -89,9 +89,9 @@ TEMP_SUDOERS="/tmp/ledmatrix_web_sudoers_$$"
     echo "$WEB_USER ALL=(ALL) NOPASSWD: $SYSTEMCTL_PATH status ledmatrix.service"
     echo "$WEB_USER ALL=(ALL) NOPASSWD: $SYSTEMCTL_PATH is-active ledmatrix"
     echo "$WEB_USER ALL=(ALL) NOPASSWD: $SYSTEMCTL_PATH is-active ledmatrix.service"
-    echo "$WEB_USER ALL=(ALL) NOPASSWD: $SYSTEMCTL_PATH start ledmatrix-web"
-    echo "$WEB_USER ALL=(ALL) NOPASSWD: $SYSTEMCTL_PATH stop ledmatrix-web"
-    echo "$WEB_USER ALL=(ALL) NOPASSWD: $SYSTEMCTL_PATH restart ledmatrix-web"
+    echo "$WEB_USER ALL=(ALL) NOPASSWD: $SYSTEMCTL_PATH start ledmatrix-web.service"
+    echo "$WEB_USER ALL=(ALL) NOPASSWD: $SYSTEMCTL_PATH stop ledmatrix-web.service"
+    echo "$WEB_USER ALL=(ALL) NOPASSWD: $SYSTEMCTL_PATH restart ledmatrix-web.service"
 
     # Optional: journalctl (non-critical — skip if not found)
     if [ -n "$JOURNALCTL_PATH" ]; then

scripts/utils/wifi_monitor_daemon.py

@@ -10,6 +10,7 @@ import sys
 import time
 import logging
 import signal
+import subprocess
 from pathlib import Path
 
 # Add project root to path (parent of scripts/utils/)
@@ -43,7 +44,11 @@ class WiFiMonitorDaemon:
         self.wifi_manager = WiFiManager()
         self.running = True
         self.last_state = None
-        
+        # Counts consecutive checks where nmcli says "connected" but internet is unreachable.
+        # After _nm_restart_threshold failures, NetworkManager is restarted as a recovery step.
+        self._consecutive_internet_failures = 0
+        self._nm_restart_threshold = 5  # ~2.5 min at 30s interval
+
         # Register signal handlers for graceful shutdown
         signal.signal(signal.SIGINT, self._signal_handler)
         signal.signal(signal.SIGTERM, self._signal_handler)
@@ -122,6 +127,43 @@ class WiFiMonitorDaemon:
                     else:
                         logger.debug(f"Status check: WiFi=disconnected, Ethernet={updated_ethernet}, AP={updated_status.ap_mode_active}")
                 
+                # Escalating recovery: if nmcli reports connected but actual internet
+                # is unreachable for several consecutive checks, restart NetworkManager.
+                # This is done HERE (not inside check_and_manage_ap_mode) to keep the
+                # AP-enable trigger clean and avoid false-positive AP enables from
+                # transient packet loss on otherwise working WiFi.
+                if updated_status.connected and not updated_status.ap_mode_active:
+                    if not self.wifi_manager.check_internet_connectivity():
+                        self._consecutive_internet_failures += 1
+                        logger.warning(
+                            f"Internet unreachable despite nmcli connection "
+                            f"({self._consecutive_internet_failures}/{self._nm_restart_threshold})"
+                        )
+                        if self._consecutive_internet_failures >= self._nm_restart_threshold:
+                            logger.warning("Restarting NetworkManager to recover internet connectivity")
+                            try:
+                                subprocess.run(
+                                    ["/usr/bin/systemctl", "restart", "NetworkManager"],
+                                    capture_output=True, timeout=20, check=True
+                                )
+                                self._consecutive_internet_failures = 0
+                                # NM restart causes a brief WiFi drop; reset the AP-mode grace
+                                # counter so that transient disconnect doesn't count toward
+                                # triggering AP mode.
+                                self.wifi_manager._disconnected_checks = 0
+                            except subprocess.CalledProcessError as e:
+                                logger.error(f"NetworkManager restart failed (rc={e.returncode}); "
+                                             "resetting failure counter to avoid tight retry loop")
+                                self._consecutive_internet_failures = 0
+                            except (subprocess.SubprocessError, OSError) as e:
+                                logger.error(f"NetworkManager restart error: {e}; "
+                                             "resetting failure counter to avoid tight retry loop")
+                                self._consecutive_internet_failures = 0
+                    else:
+                        self._consecutive_internet_failures = 0
+                else:
+                    self._consecutive_internet_failures = 0
+
                 # Sleep until next check
                 time.sleep(self.check_interval)
                 

src/cache/cache_strategy.py

@@ -193,19 +193,21 @@ class CacheStrategy:
             Data type string for strategy lookup
         """
         key_lower = key.lower()
-        
-        # Odds data — checked FIRST because odds keys may also contain 'live'/'current'
-        # (e.g. odds_espn_nba_game_123_live). The odds TTL (120s for live, 1800s for
-        # upcoming) must win over the generic sports_live TTL (30s) to avoid hitting
-        # the ESPN odds API every 30 seconds per game.
-        if 'odds' in key_lower:
-            # For live games, use shorter cache; for upcoming games, use longer cache
-            if any(x in key_lower for x in ['live', 'current']):
-                return 'odds_live'  # Live odds change more frequently (120s TTL)
-            return 'odds'  # Regular odds for upcoming games (1800s TTL)
 
-        # Live sports data (only reached if key does NOT contain 'odds')
+        # Odds data — checked before the generic 'live' block below because
+        # live-odds cache keys (e.g. odds_espn_basketball_nba_<id>_live) contain
+        # both 'odds' AND 'live'.  Without this ordering the 'live' check below
+        # would match first and return 'sports_live' (30 s TTL) instead of the
+        # correct 'odds_live' (120 s TTL).
+        if 'odds' in key_lower:
+            if any(x in key_lower for x in ['live', 'current']):
+                return 'odds_live'  # Live odds change more frequently
+            return 'odds'  # Regular odds for upcoming games
+
+        # Live sports data
         if any(x in key_lower for x in ['live', 'current', 'scoreboard']):
+            if 'soccer' in key_lower:
+                return 'sports_live'  # Soccer live data is very time-sensitive
             return 'sports_live'
 
         # Weather data

src/common/sync_manager.py

@@ -0,0 +1,651 @@
+"""
+Multi-Display Sync Manager
+
+Synchronizes scrolling content across two LED matrix display units over UDP.
+Runs at the core framework level — works with any plugin automatically.
+
+Roles:
+  standalone  No sync (default behavior)
+  leader      Drives scroll, sends rendered follower frames via UDP
+  follower    Receives frames from leader; falls back to own plugins when
+              the leader goes offline
+
+Compatibility rule: rows and cols must match between leader and follower.
+chain_length may differ — each display can have a different number of panels.
+
+Port default: 5765 (UDP). Open this port on both Pis if ufw is active:
+  sudo ufw allow 5765/udp
+"""
+
+import io
+import json
+import os
+import socket
+import struct
+import tempfile
+import threading
+import time
+import logging
+from enum import Enum
+from typing import Callable, Optional
+import numpy as np
+from PIL import Image
+
+# Raw-frame wire format: 8-byte magic + 4-byte header + raw RGB pixels
+# Much faster than PNG: no encode/decode, negligible CPU, same UDP packet size
+_RAW_MAGIC = b'SYNC_RAW'
+_RAW_HEADER = struct.Struct('<HH')  # width, height (uint16 LE)
+
+
+SYNC_PORT = 5765
+HELLO_INTERVAL = 5.0       # follower broadcasts hello every 5 s
+HEARTBEAT_INTERVAL = 2.0   # follower sends heartbeat every 2 s
+PEER_TIMEOUT = 6.0         # leader: no heartbeat → follower gone
+LEADER_TIMEOUT = 6.0       # follower: no frame → leader gone
+STATUS_FILE = os.path.join(tempfile.gettempdir(), "led_matrix_sync_status.json")
+
+
+class SyncRole(Enum):
+    STANDALONE = "standalone"
+    LEADER = "leader"
+    FOLLOWER = "follower"
+
+
+class LeaderState(Enum):
+    NO_PEER = "no_peer"
+    CONNECTED = "connected"
+    INCOMPATIBLE = "incompatible"
+
+
+class FollowerState(Enum):
+    STANDALONE = "standalone"
+    FOLLOWER = "follower"
+
+
+class DisplaySyncManager:
+    """
+    Core sync manager.  Instantiated by DisplayController based on config['sync'].
+    Leader sends compressed PNG frames to the follower after each render cycle.
+    Follower renders received frames; returns to own plugin stack when leader
+    goes offline.
+    """
+
+    def __init__(
+        self,
+        role_str: str,
+        cfg: dict,
+        hw_config: dict,
+        logger: logging.Logger,
+    ) -> None:
+        """
+        Args:
+            role_str:   "standalone" | "leader" | "follower"
+            cfg:        config['sync'] dict
+            hw_config:  config['display']['hardware'] dict (this Pi's own config)
+            logger:     framework logger
+        """
+        try:
+            self.role = SyncRole(role_str)
+        except ValueError:
+            logger.warning("Invalid sync role '%s', defaulting to standalone", role_str)
+            self.role = SyncRole.STANDALONE
+
+        self.logger = logger
+        self.port = int(cfg.get("port", SYNC_PORT))
+        self._hw_config = hw_config
+
+        # Leader state
+        self._leader_state = LeaderState.NO_PEER
+        self._peer_ip: Optional[str] = None
+        self._peer_compatible: bool = False
+        self._peer_chain: int = 0
+        self._last_heartbeat_time: float = 0.0
+        self._leader_width: int = 0  # set by display_controller after init
+
+        # Follower state
+        self._follower_state = FollowerState.STANDALONE
+        self._latest_frame: Optional[Image.Image] = None  # pixel-frame fallback
+        self._latest_scroll_x: Optional[float] = None    # Vegas scroll position
+        self._last_leader_frame_time: float = 0.0
+        self._frame_lock = threading.Lock()
+        self._leader_ip: Optional[str] = None
+        self._on_new_cycle: Optional[Callable[[], None]] = None       # called when leader starts new cycle
+        self._on_scroll_image: Optional[Callable[[Image.Image], None]] = None   # called with Image when received
+        self._pending_scroll_image: Optional[Image.Image] = None  # image received before callback set
+        self._scroll_image_lock = threading.Lock()         # guards _on_scroll_image / _pending_scroll_image
+        self._img_server_sock = None                        # TCP server for scroll image transfer
+
+        # Leader state additions
+        self._on_follower_connected: Optional[Callable[[], None]] = None  # called when follower connects
+
+        self._error_message: Optional[str] = None
+        self._running = False
+        self._recv_sock: Optional[socket.socket] = None
+        self._send_sock: Optional[socket.socket] = None
+
+        if self.role == SyncRole.STANDALONE:
+            return
+
+        if self.role == SyncRole.LEADER:
+            self._start_leader()
+        elif self.role == SyncRole.FOLLOWER:
+            self._start_follower()
+
+    # ------------------------------------------------------------------ #
+    # Leader setup                                                         #
+    # ------------------------------------------------------------------ #
+
+    def _start_leader(self) -> None:
+        # Receive socket: listens for hello + heartbeat from follower
+        self._recv_sock = socket.socket(socket.AF_INET, socket.SOCK_DGRAM)  # nosec B104
+        self._recv_sock.setsockopt(socket.SOL_SOCKET, socket.SO_REUSEADDR, 1)
+        self._recv_sock.bind(("", self.port))  # nosec B104 — intentional: must receive UDP broadcast on all interfaces
+        self._recv_sock.settimeout(1.0)
+
+        # Send socket: unicast frames + hello_ack to follower
+        self._send_sock = socket.socket(socket.AF_INET, socket.SOCK_DGRAM)
+
+        self._running = True
+        threading.Thread(
+            target=self._leader_recv_loop, daemon=True, name="sync-leader-recv"
+        ).start()
+        threading.Thread(
+            target=self._leader_watchdog, daemon=True, name="sync-leader-watchdog"
+        ).start()
+        self.logger.info("Sync: leader started on UDP port %d", self.port)
+        self.write_status_file()
+
+    def _leader_recv_loop(self) -> None:
+        while self._running:
+            try:
+                data, addr = self._recv_sock.recvfrom(1024)
+                sender_ip = addr[0]
+                try:
+                    msg = json.loads(data.decode("utf-8"))
+                except (json.JSONDecodeError, UnicodeDecodeError):
+                    continue
+                t = msg.get("t")
+                if t == "hello":
+                    self._handle_hello(msg, sender_ip)
+                elif t == "hb":
+                    if self._peer_ip == sender_ip:
+                        self._last_heartbeat_time = time.time()
+            except socket.timeout:
+                continue
+            except Exception as exc:
+                self.logger.debug("Sync leader recv error: %s", exc)
+
+    def _handle_hello(self, msg: dict, sender_ip: str) -> None:
+        hw = self._hw_config
+        local_rows = hw.get("rows", 32)
+        local_cols = hw.get("cols", 64)
+        peer_rows = int(msg.get("rows", 0))
+        peer_cols = int(msg.get("cols", 0))
+        peer_chain = int(msg.get("chain", 1))
+
+        compatible = peer_rows == local_rows and peer_cols == local_cols
+
+        self._peer_ip = sender_ip
+        self._peer_compatible = compatible
+        self._peer_chain = peer_chain
+        self._last_heartbeat_time = time.time()
+
+        prev_state = self._leader_state
+        if compatible:
+            if prev_state != LeaderState.CONNECTED:
+                self.logger.info(
+                    "Sync: follower connected at %s (chain=%d)", sender_ip, peer_chain
+                )
+            self._leader_state = LeaderState.CONNECTED
+            self._error_message = None
+            # Send scroll image immediately on new connection so follower has identical content
+            if prev_state != LeaderState.CONNECTED and self._on_follower_connected:
+                threading.Thread(
+                    target=self._on_follower_connected,
+                    daemon=True, name="sync-leader-img-push"
+                ).start()
+        else:
+            self._leader_state = LeaderState.INCOMPATIBLE
+            self._error_message = (
+                f"Incompatible panels: follower is {peer_cols}x{peer_rows}, "
+                f"leader is {local_cols}x{local_rows}. "
+                f"rows and cols must match between displays."
+            )
+            if prev_state != LeaderState.INCOMPATIBLE:
+                self.logger.error("Sync: %s", self._error_message)
+
+        if self._leader_state != prev_state:
+            self.write_status_file()
+
+        ack = json.dumps({
+            "t": "hello_ack",
+            "compatible": compatible,
+            "leader_width": self._leader_width,
+            "error": self._error_message,
+        }).encode("utf-8")
+        try:
+            self._send_sock.sendto(ack, (sender_ip, self.port))
+        except Exception as exc:
+            self.logger.debug("Sync: hello_ack send failed: %s", exc)
+
+    def _leader_watchdog(self) -> None:
+        while self._running:
+            time.sleep(1.0)
+            if self._leader_state == LeaderState.CONNECTED:
+                if time.time() - self._last_heartbeat_time > PEER_TIMEOUT:
+                    self.logger.info(
+                        "Sync: follower heartbeat timeout — peer disconnected"
+                    )
+                    self._leader_state = LeaderState.NO_PEER
+                    self._peer_ip = None
+                    self._peer_compatible = False
+                    self.write_status_file()
+
+    def _image_server_loop(self) -> None:
+        """Follower: TCP server that receives the leader's scroll image at each new cycle."""
+        while self._running:
+            try:
+                conn, addr = self._img_server_sock.accept()
+                conn.settimeout(10.0)
+                try:
+                    # 4-byte big-endian length prefix
+                    hdr = b""
+                    while len(hdr) < 4:
+                        chunk = conn.recv(4 - len(hdr))
+                        if not chunk:
+                            break
+                        hdr += chunk
+                    if len(hdr) < 4:
+                        continue
+                    length = int.from_bytes(hdr, "big")
+                    _MAX_IMAGE_BYTES = 10 * 1024 * 1024  # 10 MB — well above any real scroll image
+                    if length <= 0 or length > _MAX_IMAGE_BYTES:
+                        self.logger.warning(
+                            "Sync: rejected TCP image with invalid length %d (max %d) from %s",
+                            length, _MAX_IMAGE_BYTES, addr,
+                        )
+                        conn.close()
+                        continue
+                    data = bytearray()
+                    while len(data) < length:
+                        chunk = conn.recv(min(65536, length - len(data)))
+                        if not chunk:
+                            break
+                        data.extend(chunk)
+                    img = Image.open(io.BytesIO(data))
+                    _MAX_W, _MAX_H = 100_000, 256  # generous for any real scroll image
+                    if img.width > _MAX_W or img.height > _MAX_H:
+                        self.logger.warning(
+                            "Sync: rejected oversized scroll image %dx%d (max %dx%d) from %s",
+                            img.width, img.height, _MAX_W, _MAX_H, addr,
+                        )
+                        continue
+                    try:
+                        img.load()
+                    except (Image.DecompressionBombError, ValueError) as exc:
+                        self.logger.warning("Sync: rejected decompression bomb from %s: %s", addr, exc)
+                        continue
+                    self.logger.info(
+                        "Sync: received scroll image %dx%d (%d bytes compressed)",
+                        img.width, img.height, length,
+                    )
+                    with self._scroll_image_lock:
+                        if self._on_scroll_image:
+                            cb = self._on_scroll_image
+                        else:
+                            # Callback not registered yet (startup race) — cache it
+                            self._pending_scroll_image = img
+                            cb = None
+                    if cb:
+                        cb(img)
+                finally:
+                    conn.close()
+            except socket.timeout:
+                continue
+            except Exception as exc:
+                self.logger.debug("Sync: image server error: %s", exc)
+
+    def send_scroll_image(self, image: Image.Image) -> None:
+        """Leader: send the full scroll image to the follower via TCP.
+        PNG compression typically reduces a 5000×32 image to ~20–50KB,
+        transferring in <20ms on local WiFi. Called at new_cycle and on
+        first connection so both Pis always have identical cached_arrays.
+        """
+        if self.role != SyncRole.LEADER:
+            return
+        if self._leader_state != LeaderState.CONNECTED or not self._peer_ip:
+            return
+        try:
+            buf = io.BytesIO()
+            image.save(buf, format="PNG", optimize=True)
+            data = buf.getvalue()
+            with socket.socket(socket.AF_INET, socket.SOCK_STREAM) as sock:
+                sock.settimeout(5.0)
+                sock.connect((self._peer_ip, self.port + 1))
+                sock.sendall(len(data).to_bytes(4, "big") + data)
+            self.logger.info(
+                "Sync: sent scroll image %dx%d (%d bytes compressed)",
+                image.width, image.height, len(data),
+            )
+        except Exception as exc:
+            self.logger.debug("Sync: image send error: %s", exc)
+
+    def set_on_follower_connected(self, callback: Callable[[], None]) -> None:
+        """Leader: callback fired (in a thread) when a compatible follower first connects.
+        Use this to push the current scroll image immediately.
+        If a follower is already connected when this is called, fires right away
+        (handles the race where follower connects during leader startup).
+        """
+        self._on_follower_connected = callback
+        if self._leader_state == LeaderState.CONNECTED:
+            threading.Thread(
+                target=callback, daemon=True, name="sync-leader-img-push-late"
+            ).start()
+
+    def set_on_scroll_image(self, callback: Callable[[Image.Image], None]) -> None:
+        """Follower: callback fired with the received Image when leader sends scroll image.
+        If an image was received before this callback was registered (startup race),
+        fires immediately with that cached image.
+        """
+        with self._scroll_image_lock:
+            self._on_scroll_image = callback
+            pending = self._pending_scroll_image
+            self._pending_scroll_image = None
+        if pending is not None:
+            callback(pending)
+
+    def send_scroll_x(self, scroll_x: float) -> None:
+        """Leader (Vegas mode): broadcast scroll position instead of a pixel frame.
+        The follower renders from its own local pipeline at scroll_x - display_width.
+        ~20 bytes vs ~18KB for raw frames — eliminates all content-change artifacts.
+        """
+        if self.role != SyncRole.LEADER:
+            return
+        if self._leader_state != LeaderState.CONNECTED or not self._peer_ip:
+            return
+        try:
+            msg = json.dumps({"t": "sx", "x": round(scroll_x, 2)}).encode("utf-8")
+            self._send_sock.sendto(msg, (self._peer_ip, self.port))
+        except Exception as exc:
+            self.logger.debug("Sync: scroll_x send error: %s", exc)
+
+    def send_new_cycle(self) -> None:
+        """Leader: signal that a new scroll cycle has started so follower rebuilds its image."""
+        if self.role != SyncRole.LEADER:
+            return
+        if self._leader_state != LeaderState.CONNECTED or not self._peer_ip:
+            return
+        try:
+            self._send_sock.sendto(b'{"t":"nc"}', (self._peer_ip, self.port))
+        except Exception as exc:
+            self.logger.debug("Sync: new_cycle send error: %s", exc)
+
+    def send_frame(self, image: Image.Image) -> None:
+        """Leader: send a rendered frame to the follower as raw RGB bytes.
+        Raw format is orders of magnitude faster than PNG on Pi hardware —
+        no encode on sender, no decode on receiver.
+        Packet: 8-byte magic + 4-byte (width, height) header + raw RGB bytes.
+        """
+        if self.role != SyncRole.LEADER:
+            return
+        if self._leader_state != LeaderState.CONNECTED or not self._peer_ip:
+            return
+        try:
+            arr = np.asarray(image.convert("RGB"), dtype=np.uint8)
+            header = _RAW_MAGIC + _RAW_HEADER.pack(image.width, image.height)
+            data = header + arr.tobytes()
+            if len(data) <= 65000:
+                self._send_sock.sendto(data, (self._peer_ip, self.port))
+            elif not getattr(self, '_oversized_frame_warned', False):
+                self._oversized_frame_warned = True
+                self.logger.warning(
+                    "Sync: frame too large for UDP (%d bytes, max 65000) — "
+                    "image %dx%d will not be sent; use TCP image sync instead",
+                    len(data), image.width, image.height,
+                )
+        except Exception as exc:
+            self.logger.debug("Sync: frame send error: %s", exc)
+
+    def set_leader_width(self, width: int) -> None:
+        """Called by DisplayController once display_manager.width is known."""
+        self._leader_width = width
+
+    # ------------------------------------------------------------------ #
+    # Follower setup                                                       #
+    # ------------------------------------------------------------------ #
+
+    def _start_follower(self) -> None:
+        # Receive socket: listens for frames + hello_ack from leader
+        self._recv_sock = socket.socket(socket.AF_INET, socket.SOCK_DGRAM)
+        self._recv_sock.setsockopt(socket.SOL_SOCKET, socket.SO_REUSEADDR, 1)
+        self._recv_sock.bind(("", self.port))  # nosec B104 — intentional: must receive UDP broadcast on all interfaces
+        self._recv_sock.settimeout(0.1)
+
+        # Send socket: broadcasts hello + heartbeat
+        self._send_sock = socket.socket(socket.AF_INET, socket.SOCK_DGRAM)
+        self._send_sock.setsockopt(socket.SOL_SOCKET, socket.SO_BROADCAST, 1)
+
+        self._running = True
+        threading.Thread(
+            target=self._follower_recv_loop, daemon=True, name="sync-follower-recv"
+        ).start()
+        threading.Thread(
+            target=self._follower_announce_loop, daemon=True, name="sync-follower-announce"
+        ).start()
+        threading.Thread(
+            target=self._follower_watchdog, daemon=True, name="sync-follower-watchdog"
+        ).start()
+        # TCP server: receives scroll images from leader (port + 1)
+        self._img_server_sock = socket.socket(socket.AF_INET, socket.SOCK_STREAM)  # nosec B104
+        self._img_server_sock.setsockopt(socket.SOL_SOCKET, socket.SO_REUSEADDR, 1)
+        self._img_server_sock.bind(("", self.port + 1))  # nosec B104 — intentional: TCP server must accept connections on all interfaces
+        self._img_server_sock.listen(1)
+        self._img_server_sock.settimeout(1.0)
+        threading.Thread(
+            target=self._image_server_loop, daemon=True, name="sync-image-server"
+        ).start()
+
+        self.logger.info(
+            "Sync: follower started on UDP port %d, image server on TCP %d",
+            self.port, self.port + 1,
+        )
+        self.write_status_file()
+
+    def _follower_recv_loop(self) -> None:
+        while self._running:
+            try:
+                data, addr = self._recv_sock.recvfrom(65535)
+                sender_ip = addr[0]
+
+                if data[:8] == _RAW_MAGIC or len(data) > 512:
+                    # Frame data: prefer magic-tagged raw RGB; fall back to legacy PNG
+                    try:
+                        if data[:8] == _RAW_MAGIC:
+                            w, h = _RAW_HEADER.unpack(data[8:12])
+                            raw = data[12:]
+                            img = Image.frombuffer(
+                                "RGB", (w, h), raw, "raw", "RGB", 0, 1
+                            )
+                        else:
+                            # Fallback: try legacy PNG
+                            img = Image.open(io.BytesIO(data))
+                            img.load()
+                        with self._frame_lock:
+                            self._latest_frame = img
+                        self._last_leader_frame_time = time.time()
+                        self._leader_ip = sender_ip
+
+                        if self._follower_state == FollowerState.STANDALONE:
+                            self._follower_state = FollowerState.FOLLOWER
+                            self.logger.info(
+                                "Sync: leader active at %s — switching to follower mode",
+                                sender_ip,
+                            )
+                            self.write_status_file()
+                    except Exception as exc:
+                        self.logger.debug("Sync: frame decode error: %s", exc)
+                else:
+                    # Control message
+                    try:
+                        msg = json.loads(data.decode("utf-8"))
+                        t = msg.get("t")
+                        if t == "hello_ack":
+                            self._leader_ip = sender_ip
+                            self._peer_compatible = msg.get("compatible", False)
+                            self._error_message = msg.get("error")
+                            if not self._peer_compatible and self._error_message:
+                                self.logger.error(
+                                    "Sync: leader rejected handshake — %s",
+                                    self._error_message,
+                                )
+                            self.write_status_file()
+                        elif t == "sx":
+                            # Vegas scroll-position sync — tiny message, renders locally
+                            self._latest_scroll_x = float(msg["x"])
+                            self._last_leader_frame_time = time.time()
+                            self._leader_ip = sender_ip
+                            if self._follower_state == FollowerState.STANDALONE:
+                                self._follower_state = FollowerState.FOLLOWER
+                                self.logger.info(
+                                    "Sync: leader active at %s — switching to follower mode",
+                                    sender_ip,
+                                )
+                                self.write_status_file()
+                                if self._on_new_cycle:
+                                    self._on_new_cycle()  # build initial scroll image
+                        elif t == "nc":
+                            # Leader started a new scroll cycle — rebuild local image
+                            if self._on_new_cycle:
+                                self._on_new_cycle()
+                    except (json.JSONDecodeError, UnicodeDecodeError, KeyError):
+                        pass
+
+            except socket.timeout:
+                continue
+            except Exception as exc:
+                self.logger.debug("Sync follower recv error: %s", exc)
+
+    def _follower_announce_loop(self) -> None:
+        hw = self._hw_config
+        hello = json.dumps({
+            "t": "hello",
+            "rows": hw.get("rows", 32),
+            "cols": hw.get("cols", 64),
+            "chain": hw.get("chain_length", 1),
+        }).encode("utf-8")
+        heartbeat = json.dumps({"t": "hb"}).encode("utf-8")
+        dest = ("<broadcast>", self.port)
+
+        last_hello = 0.0
+        last_hb = 0.0
+
+        while self._running:
+            now = time.time()
+            if now - last_hello >= HELLO_INTERVAL:
+                try:
+                    self._send_sock.sendto(hello, dest)
+                    last_hello = now
+                except Exception as exc:
+                    self.logger.debug("Sync: hello broadcast error: %s", exc)
+            if now - last_hb >= HEARTBEAT_INTERVAL:
+                try:
+                    self._send_sock.sendto(heartbeat, dest)
+                    last_hb = now
+                except Exception as exc:
+                    self.logger.debug("Sync: heartbeat error: %s", exc)
+            time.sleep(0.5)
+
+    def _follower_watchdog(self) -> None:
+        while self._running:
+            time.sleep(1.0)
+            if self._follower_state == FollowerState.FOLLOWER:
+                if time.time() - self._last_leader_frame_time > LEADER_TIMEOUT:
+                    self.logger.info(
+                        "Sync: leader frame timeout — returning to standalone mode"
+                    )
+                    self._follower_state = FollowerState.STANDALONE
+                    with self._frame_lock:
+                        self._latest_frame = None
+                    self.write_status_file()
+
+    # ------------------------------------------------------------------ #
+    # Public API                                                           #
+    # ------------------------------------------------------------------ #
+
+    def is_follower_active(self) -> bool:
+        """True when this Pi is in active follower mode (receiving frames)."""
+        return (
+            self.role == SyncRole.FOLLOWER
+            and self._follower_state == FollowerState.FOLLOWER
+        )
+
+    def get_latest_scroll_x(self) -> Optional[float]:
+        """Follower: return the most recently received Vegas scroll position, or None."""
+        return self._latest_scroll_x
+
+    def set_on_new_cycle(self, callback: Callable[[], None]) -> None:
+        """Follower: register a callback fired when the leader starts a new scroll cycle.
+        Used to trigger a local start_new_cycle() so both Pis rebuild from same fresh data.
+        """
+        self._on_new_cycle = callback
+
+    def get_latest_frame(self) -> Optional[Image.Image]:
+        """Follower: return the most recently received pixel frame (non-Vegas fallback)."""
+        with self._frame_lock:
+            return self._latest_frame
+
+    def get_status(self) -> dict:
+        """Return sync state dict for the web API status endpoint."""
+        hw = self._hw_config
+        base = {
+            "role": self.role.value,
+            "port": self.port,
+            "local_rows": hw.get("rows", 32),
+            "local_cols": hw.get("cols", 64),
+            "local_chain": hw.get("chain_length", 1),
+        }
+
+        if self.role == SyncRole.STANDALONE:
+            return {**base, "state": "standalone"}
+
+        if self.role == SyncRole.LEADER:
+            return {
+                **base,
+                "state": self._leader_state.value,
+                "peer_ip": self._peer_ip,
+                "peer_compatible": self._peer_compatible,
+                "peer_chain": self._peer_chain,
+                "leader_width": self._leader_width,
+                "error": self._error_message,
+            }
+
+        # Follower
+        return {
+            **base,
+            "state": self._follower_state.value,
+            "leader_ip": self._leader_ip,
+            "peer_compatible": self._peer_compatible,
+            "error": self._error_message,
+        }
+
+    def write_status_file(self) -> None:
+        """Write current sync status to STATUS_FILE for the web UI to read."""
+        try:
+            status = self.get_status()
+            status["ts"] = time.time()
+            tmp = STATUS_FILE + ".tmp"
+            with open(tmp, "w") as f:
+                json.dump(status, f)
+            os.replace(tmp, STATUS_FILE)
+        except Exception as exc:
+            self.logger.debug("Sync: status file write error: %s", exc)
+
+    def stop(self) -> None:
+        """Shut down threads and close sockets."""
+        self._running = False
+        for sock in (self._recv_sock, self._send_sock, self._img_server_sock):
+            if sock:
+                try:
+                    sock.close()
+                except Exception as exc:
+                    self.logger.debug("Sync: error closing socket: %s", exc)

src/display_controller.py

@@ -16,6 +16,7 @@ from src.config_service import ConfigService
 from src.cache_manager import CacheManager
 from src.font_manager import FontManager
 from src.logging_config import get_logger
+from src.common.sync_manager import DisplaySyncManager, SyncRole
 
 # Get logger with consistent configuration
 logger = get_logger(__name__)
@@ -32,10 +33,7 @@ class DisplayController:
     def __init__(self):
         start_time = time.time()
         logger.info("Starting DisplayController initialization")
-
-        # Throttle tracking for _tick_plugin_updates in high-FPS loops
-        self._last_plugin_tick_time = 0.0
-
+        
         # Initialize ConfigManager and wrap with ConfigService for hot-reload
         config_manager = ConfigManager()
         enable_hot_reload = os.environ.get('LEDMATRIX_HOT_RELOAD', 'true').lower() == 'true'
@@ -69,7 +67,39 @@ class DisplayController:
         config_time = time.time()
         self.display_manager = DisplayManager(self.config)
         logger.info("DisplayManager initialized in %.3f seconds", time.time() - config_time)
-        
+
+        # Initialize multi-display sync (standalone by default — no-op unless configured)
+        sync_cfg = self.config.get("sync", {})
+        hw_cfg = self.config.get("display", {}).get("hardware", {})
+        self.sync_manager = DisplaySyncManager(
+            role_str=sync_cfg.get("role", "standalone"),
+            cfg=sync_cfg,
+            hw_config=hw_cfg,
+            logger=logger,
+        )
+        # Tell the leader its own physical display width so it can include it in hello_ack
+        if self.sync_manager.role == SyncRole.LEADER:
+            self.sync_manager.set_leader_width(self.display_manager.width)
+
+        # Follower mode setup
+        if self.sync_manager.role == SyncRole.FOLLOWER:
+            # Gate update_display() so background plugin threads cannot write to
+            # hardware — only our render loop is permitted.
+            _real_update = self.display_manager.update_display
+            _dm = self.display_manager
+            def _follower_gated_update():
+                # Allow through when the sync render loop has the token, or when
+                # the leader has gone offline and we've fallen back to standalone.
+                if getattr(_dm, '_sync_render_allowed', False) or not self.sync_manager.is_follower_active():
+                    _real_update()
+            self.display_manager.update_display = _follower_gated_update
+
+            # Note: _on_new_cycle is NOT registered here. The leader now sends
+            # its actual scroll image via TCP at each new_cycle, so the follower
+            # adopts that image directly via set_on_scroll_image(). Registering
+            # _on_new_cycle would trigger a local rebuild that overwrites the
+            # leader's just-received image with a different locally-built one.
+
         # Initialize Font Manager
         font_time = time.time()
         self.font_manager = FontManager(self.config)
@@ -82,8 +112,7 @@ class DisplayController:
         logger.info("Display modes initialized in %.3f seconds", time.time() - init_time)
         
         self.force_change = False
-        self._next_live_priority_check = 0.0  # monotonic timestamp for throttled live priority checks
-
+        
         # All sports and content managers now handled via plugins
         logger.info("All sports and content managers now handled via plugin system")
         
@@ -396,20 +425,64 @@ class DisplayController:
             # Set up live priority checker
             self.vegas_coordinator.set_live_priority_checker(self._check_live_priority)
 
-            # Set up interrupt checker for on-demand/wifi status
+            # Set up interrupt checker for on-demand/wifi status and follower mode
+            def _vegas_interrupt():
+                return self._check_vegas_interrupt() or self.sync_manager.is_follower_active()
             self.vegas_coordinator.set_interrupt_checker(
-                self._check_vegas_interrupt,
+                _vegas_interrupt,
                 check_interval=10  # Check every 10 frames (~80ms at 125 FPS)
             )
 
-            # Set up plugin update tick to keep data fresh during Vegas mode
-            self.vegas_coordinator.set_update_tick(
-                self._tick_plugin_updates_for_vegas,
-                interval=1.0
-            )
+            # Run plugin updates inside the Vegas loop so the inter-iteration
+            # gap is <1 ms (nothing left for _tick_plugin_updates() to do).
+            self.vegas_coordinator.set_update_callback(self._tick_plugin_updates)
+
+            # Wire multi-display sync into Vegas render pipeline
+            follower_pos = self.config.get("sync", {}).get("follower_position", "left")
+            self.vegas_coordinator.set_sync_manager(self.sync_manager, follower_pos)
 
             logger.info("Vegas mode coordinator initialized")
 
+            # Follower does NOT build its own initial scroll image — the leader
+            # pushes its image via TCP as soon as set_on_follower_connected fires.
+            # A local build would create a different (wrong) image that could
+            # temporarily replace the leader's correct one.
+
+            # When the leader sends its scroll image (TCP), update our
+            # cached_array so both Pis have pixel-identical images.
+            import numpy as _np
+            def _on_leader_scroll_image(image):
+                vc = getattr(self, 'vegas_coordinator', None)
+                if vc and vc.render_pipeline:
+                    rp = vc.render_pipeline
+                    arr = _np.asarray(image.convert("RGB"), dtype=_np.uint8)
+                    rp.scroll_helper.cached_image = image
+                    rp.scroll_helper.cached_array = arr
+                    rp.scroll_helper.total_scroll_width = image.width
+                    self._follower_pending_new_image = False
+                    logger.info(
+                        "Sync: follower adopted leader scroll image %dx%d",
+                        image.width, image.height,
+                    )
+            self.sync_manager.set_on_scroll_image(_on_leader_scroll_image)
+
+            if self.sync_manager.role == SyncRole.LEADER:
+                # When a follower first connects, push the current scroll image so
+                # the follower doesn't have to wait for the next new_cycle event.
+                # Polls until the image is ready (Vegas may still be composing on startup).
+                def _on_follower_connected():
+                    import time as _t
+                    for _ in range(300):  # up to 30s
+                        vc = getattr(self, 'vegas_coordinator', None)
+                        if vc and vc.render_pipeline:
+                            img = vc.render_pipeline.scroll_helper.cached_image
+                            if img is not None:
+                                self.sync_manager.send_scroll_image(img)
+                                return
+                        _t.sleep(0.1)
+                    logger.warning("Sync: no scroll image available to push to new follower")
+                self.sync_manager.set_on_follower_connected(_on_follower_connected)
+
         except Exception as e:
             logger.error("Failed to initialize Vegas mode: %s", e, exc_info=True)
             self.vegas_coordinator = None
@@ -444,51 +517,16 @@ class DisplayController:
 
         return False
 
-    def _tick_plugin_updates_for_vegas(self):
-        """
-        Run scheduled plugin updates and return IDs of plugins that were updated.
-
-        Called periodically by the Vegas coordinator to keep plugin data fresh
-        during Vegas mode. Returns a list of plugin IDs whose data changed so
-        Vegas can refresh their content in the scroll.
-
-        Returns:
-            List of updated plugin IDs, or None if no updates occurred
-        """
-        if not self.plugin_manager or not hasattr(self.plugin_manager, 'plugin_last_update'):
-            self._tick_plugin_updates()
-            return None
-
-        # Snapshot update timestamps before ticking
-        old_times = dict(self.plugin_manager.plugin_last_update)
-
-        # Run the scheduled updates
-        self._tick_plugin_updates()
-
-        # Detect which plugins were actually updated
-        updated = []
-        for plugin_id, new_time in self.plugin_manager.plugin_last_update.items():
-            if new_time > old_times.get(plugin_id, 0.0):
-                updated.append(plugin_id)
-
-        if updated:
-            logger.info("Vegas update tick: %d plugin(s) updated: %s", len(updated), updated)
-
-        return updated or None
-
     def _check_schedule(self):
         """Check if display should be active based on schedule."""
-        # Get fresh config from config_service to support hot-reload
-        current_config = self.config_service.get_config()
-
-        schedule_config = current_config.get('schedule', {})
-
+        schedule_config = self.config.get('schedule', {})
+        
         # If schedule config doesn't exist or is empty, default to always active
         if not schedule_config:
             self.is_display_active = True
             self._was_display_active = True  # Track previous state for schedule change detection
             return
-
+        
         # Check if schedule is explicitly disabled
         # Default to True (schedule enabled) if 'enabled' key is missing for backward compatibility
         if 'enabled' in schedule_config and not schedule_config.get('enabled', True):
@@ -498,7 +536,7 @@ class DisplayController:
             return
 
         # Get configured timezone, default to UTC
-        timezone_str = current_config.get('timezone', 'UTC')
+        timezone_str = self.config.get('timezone', 'UTC')
         try:
             tz = pytz.timezone(timezone_str)
         except pytz.UnknownTimeZoneError:
@@ -596,18 +634,15 @@ class DisplayController:
             Target brightness level (dim_brightness if in dim period,
             normal brightness otherwise)
         """
-        # Get fresh config from config_service to support hot-reload
-        current_config = self.config_service.get_config()
-
         # Get normal brightness from config
-        normal_brightness = current_config.get('display', {}).get('hardware', {}).get('brightness', 90)
+        normal_brightness = self.config.get('display', {}).get('hardware', {}).get('brightness', 90)
 
         # If display is OFF via schedule, don't process dim schedule
         if not self.is_display_active:
             self.is_dimmed = False
             return normal_brightness
 
-        dim_config = current_config.get('dim_schedule', {})
+        dim_config = self.config.get('dim_schedule', {})
 
         # If dim schedule doesn't exist or is disabled, use normal brightness
         if not dim_config or not dim_config.get('enabled', False):
@@ -615,7 +650,7 @@ class DisplayController:
             return normal_brightness
 
         # Get configured timezone
-        timezone_str = current_config.get('timezone', 'UTC')
+        timezone_str = self.config.get('timezone', 'UTC')
         try:
             tz = pytz.timezone(timezone_str)
         except pytz.UnknownTimeZoneError:
@@ -722,21 +757,83 @@ class DisplayController:
             except Exception:  # pylint: disable=broad-except
                 logger.exception("Error running scheduled plugin updates")
 
-    def _tick_plugin_updates_throttled(self, min_interval: float = 0.0):
-        """Throttled version of _tick_plugin_updates for high-FPS loops.
+    _FOLLOWER_SEND_INTERVAL = 1.0 / 90  # raw bytes are cheap; 90fps > follower render rate
 
-        Args:
-            min_interval: Minimum seconds between calls.  When <= 0 the
-                call passes straight through to _tick_plugin_updates so
-                plugin-configured update_interval values are never capped.
+    def _follower_rebuild_scroll_image(self) -> None:
+        """Follower: rebuild the local Vegas scroll image so both Pis render from
+        the same fresh plugin data. Called at startup (after Vegas initializes)
+        and each time the leader broadcasts a new-cycle signal. Runs in a daemon
+        thread so it never blocks the 60fps render loop.
         """
-        if min_interval <= 0:
-            self._tick_plugin_updates()
+        try:
+            vc = getattr(self, 'vegas_coordinator', None)
+            if not vc:
+                logger.warning("Sync: follower has no vegas_coordinator — cannot build scroll image")
+                return
+            rp = vc.render_pipeline
+            if not rp:
+                logger.warning("Sync: follower vegas_coordinator has no render_pipeline")
+                return
+            logger.info("Sync: follower starting scroll image rebuild")
+            ok = rp.start_new_cycle()
+            if ok and rp.scroll_helper.cached_image is not None:
+                logger.info(
+                    "Sync: follower scroll image ready — %dx%d",
+                    rp.scroll_helper.cached_image.width,
+                    rp.scroll_helper.cached_image.height,
+                )
+            else:
+                logger.warning(
+                    "Sync: follower scroll image rebuild FAILED (ok=%s, cached=%s)",
+                    ok, rp.scroll_helper.cached_image is not None,
+                )
+        except Exception as exc:
+            logger.warning("Sync: follower scroll image rebuild error: %s", exc, exc_info=True)
+
+    def _send_follower_frame(self, plugin_instance) -> None:
+        """Leader: generate and send the follower's portion of the current frame.
+
+        The follower is physically to the LEFT of the leader in a right-to-left
+        scrolling ticker, so it shows content at scroll_position - display_width
+        (content that already scrolled off the leader's left edge).
+        Set sync.follower_position = "right" in config to invert this.
+        """
+        if not (self.sync_manager and self.sync_manager.role == SyncRole.LEADER):
             return
+        # Throttle to ~90fps via _FOLLOWER_SEND_INTERVAL — raw RGB bytes, no encode/decode
         now = time.time()
-        if now - self._last_plugin_tick_time >= min_interval:
-            self._last_plugin_tick_time = now
-            self._tick_plugin_updates()
+        if now - getattr(self, '_last_follower_send', 0) < self._FOLLOWER_SEND_INTERVAL:
+            return
+        self._last_follower_send = now
+
+        follower_frame = None
+        width = self.display_manager.width
+        sync_cfg = self.config.get("sync", {})
+        sign = -1 if sync_cfg.get("follower_position", "left") == "left" else 1
+        offset = sign * width
+
+        # 1. Explicit hook — plugin opted in with get_offset_frame()
+        try:
+            follower_frame = plugin_instance.get_offset_frame(offset)
+        except Exception:
+            pass
+
+        # 2. Auto-detect — plugin has a scroll_helper (standard pattern for all
+        #    scroll plugins). Works with zero plugin code changes.
+        if follower_frame is None:
+            try:
+                scroll_h = getattr(plugin_instance, 'scroll_helper', None)
+                if scroll_h is not None:
+                    follower_frame = scroll_h.get_portion_at(scroll_h.scroll_position + offset)
+            except Exception:
+                pass
+
+        # 3. Mirror fallback — static plugins (clock, weather) show same frame
+        if follower_frame is None:
+            follower_frame = self.display_manager.image
+
+        if follower_frame is not None:
+            self.sync_manager.send_frame(follower_frame)
 
     def _sleep_with_plugin_updates(self, duration: float, tick_interval: float = 1.0):
         """Sleep while continuing to service plugin update schedules."""
@@ -1373,6 +1470,88 @@ class DisplayController:
                 # Plugins update on their own schedules - no forced sync updates needed
                 # Each plugin has its own update_interval and background services
                 
+                # Multi-display sync: follower mode — render frames received from leader.
+                # Plugin update() threads still run (via _tick_plugin_updates above) so
+                # data is fresh when we return to standalone if the leader goes offline.
+                if self.sync_manager.is_follower_active():
+                    # Dead-reckoning follower render:
+                    # Advance local position at configured speed each tick; snap or
+                    # gently correct toward received scroll_x to absorb UDP jitter.
+                    _now_dr = time.perf_counter()
+                    _dt = _now_dr - getattr(self, '_follower_dr_last_t', _now_dr)
+                    self._follower_dr_last_t = _now_dr
+
+                    vc = getattr(self, 'vegas_coordinator', None)
+                    rp = vc.render_pipeline if (vc and vc.render_pipeline) else None
+                    width = self.display_manager.width
+
+                    # Advance local position at Vegas scroll speed (px/s → px/tick)
+                    vegas_speed = (
+                        self.config.get('display', {})
+                            .get('vegas_scroll', {})
+                            .get('scroll_speed', 75)
+                    )
+                    local_x = getattr(self, '_follower_local_x', None)
+                    if local_x is None:
+                        local_x = float(width)  # safe start (past pre-roll guard)
+                    local_x += vegas_speed * _dt
+
+                    # Pull latest position from leader (may be None if no packet yet)
+                    scroll_x = self.sync_manager.get_latest_scroll_x()
+                    if scroll_x is not None:
+                        diff = scroll_x - local_x
+                        total_w = (
+                            rp.scroll_helper.total_scroll_width
+                            if rp and rp.scroll_helper.total_scroll_width
+                            else width * 4
+                        )
+                        if abs(diff) > total_w * 0.5:
+                            # Large jump → cycle reset, snap immediately
+                            local_x = float(scroll_x)
+                            self._follower_pending_new_image = True
+                        elif abs(diff) > 10:
+                            # Moderate drift → 20% correction per tick
+                            local_x += diff * 0.20
+                        else:
+                            # Near → gentle 5% correction
+                            local_x += diff * 0.05
+
+                    self._follower_local_x = local_x
+
+                    if rp and rp.scroll_helper.cached_image is not None:
+                        sync_cfg = self.config.get("sync", {})
+                        sign = -1 if sync_cfg.get("follower_position", "left") == "left" else 1
+                        # Hold last frame until TCP image arrives after cycle reset
+                        if not getattr(self, "_follower_pending_new_image", False):
+                            if local_x >= width:
+                                rp.scroll_helper.scroll_position = local_x + sign * width
+                                frame = rp.scroll_helper.get_visible_portion()
+                                if frame is not None:
+                                    self._follower_last_frame = frame
+                    elif scroll_x is None:
+                        # Fallback: pixel frame before first scroll_x arrives
+                        frame = self.sync_manager.get_latest_frame()
+                        if frame is not None:
+                            self._follower_last_frame = frame
+
+                    display_frame = getattr(self, '_follower_last_frame', None)
+                    if display_frame is not None:
+                        self.display_manager.image = display_frame
+                        self.display_manager._sync_render_allowed = True
+                        self.display_manager.update_display()
+                        self.display_manager._sync_render_allowed = False
+                    # Precision deadline timer — keeps render at exactly 60fps
+                    _deadline = getattr(self, '_follower_deadline', None)
+                    _now = time.perf_counter()
+                    if _deadline is None or _now > _deadline + 0.1:
+                        _deadline = _now
+                    _deadline += 1.0 / 60
+                    self._follower_deadline = _deadline
+                    _sleep = _deadline - time.perf_counter()
+                    if _sleep > 0:
+                        time.sleep(_sleep)
+                    continue
+
                 # Process any deferred updates that may have accumulated
                 # This also cleans up expired updates to prevent memory leaks
                 self.display_manager.process_deferred_updates()
@@ -1746,7 +1925,7 @@ class DisplayController:
                             )
 
                         target_duration = max_duration
-                        start_time = time.monotonic()
+                        start_time = time.time()
 
                         def _should_exit_dynamic(elapsed_time: float) -> bool:
                             if not dynamic_enabled:
@@ -1805,34 +1984,19 @@ class DisplayController:
                                 except Exception:  # pylint: disable=broad-except
                                     logger.exception("Error during display update")
 
+                                # Multi-display sync: send follower frame after each render
+                                self._send_follower_frame(manager_to_display)
+
                                 time.sleep(display_interval)
-                                self._tick_plugin_updates_throttled(min_interval=1.0)
+                                self._tick_plugin_updates()
                                 self._poll_on_demand_requests()
                                 self._check_on_demand_expiration()
 
-                                # Check for live priority every ~30s so live
-                                # games can interrupt long display durations
-                                elapsed = time.monotonic() - start_time
-                                now = time.monotonic()
-                                if not self.on_demand_active and now >= self._next_live_priority_check:
-                                    self._next_live_priority_check = now + 30.0
-                                    live_mode = self._check_live_priority()
-                                    if live_mode and live_mode != active_mode:
-                                        logger.info("Live priority detected during high-FPS loop: %s", live_mode)
-                                        self.current_display_mode = live_mode
-                                        self.force_change = True
-                                        try:
-                                            self.current_mode_index = self.available_modes.index(live_mode)
-                                        except ValueError:
-                                            pass
-                                        # continue the main while loop to skip
-                                        # post-loop rotation/sleep logic
-                                        break
-
                                 if self.current_display_mode != active_mode:
                                     logger.debug("Mode changed during high-FPS loop, breaking early")
                                     break
 
+                                elapsed = time.time() - start_time
                                 if elapsed >= target_duration:
                                     logger.debug(
                                         "Reached high-FPS target duration %.2fs for mode %s",
@@ -1862,7 +2026,7 @@ class DisplayController:
                                 time.sleep(display_interval)
                                 self._tick_plugin_updates()
 
-                                elapsed = time.monotonic() - start_time
+                                elapsed = time.time() - start_time
                                 if elapsed >= target_duration:
                                     logger.debug(
                                         "Reached standard target duration %.2fs for mode %s",
@@ -1889,25 +2053,11 @@ class DisplayController:
                                 except Exception:  # pylint: disable=broad-except
                                     logger.exception("Error during display update")
 
+                                # Multi-display sync: send follower frame after each render
+                                self._send_follower_frame(manager_to_display)
+
                                 self._poll_on_demand_requests()
                                 self._check_on_demand_expiration()
-
-                                # Check for live priority every ~30s so live
-                                # games can interrupt long display durations
-                                now = time.monotonic()
-                                if not self.on_demand_active and now >= self._next_live_priority_check:
-                                    self._next_live_priority_check = now + 30.0
-                                    live_mode = self._check_live_priority()
-                                    if live_mode and live_mode != active_mode:
-                                        logger.info("Live priority detected during display loop: %s", live_mode)
-                                        self.current_display_mode = live_mode
-                                        self.force_change = True
-                                        try:
-                                            self.current_mode_index = self.available_modes.index(live_mode)
-                                        except ValueError:
-                                            pass
-                                        break
-
                                 if self.current_display_mode != active_mode:
                                     logger.info("Mode changed during display loop from %s to %s, breaking early", active_mode, self.current_display_mode)
                                     break
@@ -1921,26 +2071,19 @@ class DisplayController:
                                     loop_completed = True
                                     break
 
-                        # If live priority preempted the display loop, skip
-                        # all post-loop logic (remaining sleep, rotation) and
-                        # restart the main loop so the live mode displays
-                        # immediately.
-                        if self.current_display_mode != active_mode:
-                            continue
-
                         # Ensure we honour minimum duration when not dynamic and loop ended early
                         if (
                             not dynamic_enabled
                             and not loop_completed
                             and not needs_high_fps
                         ):
-                            elapsed = time.monotonic() - start_time
+                            elapsed = time.time() - start_time
                             remaining_sleep = max(0.0, max_duration - elapsed)
                             if remaining_sleep > 0:
                                 self._sleep_with_plugin_updates(remaining_sleep)
 
                         if dynamic_enabled:
-                            elapsed_total = time.monotonic() - start_time
+                            elapsed_total = time.time() - start_time
                             cycle_done = self._plugin_cycle_complete(manager_to_display)
                             
                             # Log cycle completion status and metrics

src/display_manager.py

@@ -3,6 +3,7 @@ if os.getenv("EMULATOR", "false") == "true":
     from RGBMatrixEmulator import RGBMatrix, RGBMatrixOptions
 else:
     from rgbmatrix import RGBMatrix, RGBMatrixOptions
+from contextlib import contextmanager
 from PIL import Image, ImageDraw, ImageFont
 import time
 from typing import Dict, Any, List, Tuple
@@ -28,6 +29,8 @@ class DisplayManager:
         self.config = config or {}
         self._force_fallback = force_fallback
         self._suppress_test_pattern = suppress_test_pattern
+        # When True, update_display() and clear() skip hardware writes (used during off-screen content capture)
+        self._capture_mode_active = False
         # Snapshot settings for web preview integration (service writes, web reads)
         self._snapshot_path = "/tmp/led_matrix_preview.png"
         self._snapshot_min_interval_sec = 0.2  # max ~5 fps
@@ -255,6 +258,22 @@ class DisplayManager:
         except Exception as e:
             logger.error(f"Error drawing test pattern: {e}", exc_info=True)
 
+    @contextmanager
+    def capture_mode(self):
+        """Suppress hardware output during off-screen content capture.
+
+        Plugins call update_display() as part of their normal display() flow.
+        When fetching content for Vegas mode the render loop is still running,
+        so any incidental hardware write causes a visible flash on the matrix.
+        Entering this context prevents those writes without affecting the PIL
+        image buffer, which the adapter reads to extract content.
+        """
+        self._capture_mode_active = True
+        try:
+            yield
+        finally:
+            self._capture_mode_active = False
+
     def update_display(self):
         """Update the display using double buffering with proper sync."""
         try:
@@ -264,10 +283,13 @@ class DisplayManager:
                 # Still write a snapshot so the web UI can preview
                 self._write_snapshot_if_due()
                 return
-                
-            # Copy the current image to the offscreen canvas   
+
+            if self._capture_mode_active:
+                return  # Skip hardware write — content is being captured off-screen
+
+            # Copy the current image to the offscreen canvas
             self.offscreen_canvas.SetImage(self.image)
-            
+
             # Swap buffers immediately
             self.matrix.SwapOnVSync(self.offscreen_canvas)
             
@@ -304,21 +326,23 @@ class DisplayManager:
             # Create a new black image
             self.image = Image.new('RGB', (self.matrix.width, self.matrix.height))
             self.draw = ImageDraw.Draw(self.image)
-            
-            # Clear both canvases and the underlying matrix to ensure no artifacts
-            try:
-                self.offscreen_canvas.Clear()
-            except Exception:
-                pass
-            try:
-                self.current_canvas.Clear()
-            except Exception:
-                pass
-            try:
-                # Extra safety: clear the matrix front buffer as well
-                self.matrix.Clear()
-            except Exception:
-                pass
+
+            if not self._capture_mode_active:
+                # Clear both canvases and the underlying matrix to ensure no artifacts.
+                # Failures are non-fatal — the image buffer is already black above, so
+                # the next update_display() call will push clean content regardless.
+                try:
+                    self.offscreen_canvas.Clear()
+                except (RuntimeError, OSError) as e:
+                    logger.error("Failed to clear offscreen canvas: %s", e)
+                try:
+                    self.current_canvas.Clear()
+                except (RuntimeError, OSError) as e:
+                    logger.error("Failed to clear current canvas: %s", e)
+                try:
+                    self.matrix.Clear()
+                except (RuntimeError, OSError) as e:
+                    logger.error("Failed to clear matrix front buffer: %s", e)
             
             # Note: We do NOT call update_display() here to avoid black flashes.
             # The caller should call update_display() after drawing new content.

src/plugin_system/store_manager.py

@@ -1850,58 +1850,72 @@ class PluginStoreManager:
                 return cached[1]
 
         try:
-            sha_result = subprocess.run(
-                ['git', '-C', str(plugin_path), 'rev-parse', 'HEAD'],
+            # .git may be a file (worktree / submodule) containing "gitdir: <path>".
+            # Resolve it to the actual git directory before reading any files.
+            try:
+                if git_dir.is_file():
+                    pointer = git_dir.read_text(encoding='utf-8', errors='replace').strip()
+                    if pointer.startswith('gitdir:'):
+                        resolved = (plugin_path / pointer[len('gitdir:'):].strip()).resolve()
+                        if resolved.is_dir():
+                            git_dir = resolved
+                        else:
+                            return None
+                    else:
+                        return None
+            except (OSError, NotADirectoryError):
+                return None
+
+            # Read branch directly from .git/HEAD (no subprocess).
+            branch = ''
+            try:
+                head_text = (git_dir / 'HEAD').read_text(encoding='utf-8', errors='replace').strip()
+                if head_text.startswith('ref: refs/heads/'):
+                    branch = head_text[len('ref: refs/heads/'):]
+                elif head_text.startswith('ref: '):
+                    branch = head_text[len('ref: '):]
+                # else: detached HEAD — branch stays ''
+            except (OSError, NotADirectoryError):
+                pass
+
+            # Remote URL from .git/config — parse [remote "origin"] url line.
+            remote_url = None
+            try:
+                config_text = (git_dir / 'config').read_text(encoding='utf-8', errors='replace')
+                in_origin = False
+                for line in config_text.splitlines():
+                    stripped = line.strip()
+                    if stripped == '[remote "origin"]':
+                        in_origin = True
+                    elif stripped.startswith('['):
+                        in_origin = False
+                    elif in_origin and stripped.startswith('url') and '=' in stripped:
+                        remote_url = stripped.split('=', 1)[1].strip()
+                        break
+            except (OSError, NotADirectoryError):
+                pass
+
+            # Single subprocess: SHA + commit date in one call.
+            log_result = subprocess.run(
+                ['git', '-C', str(plugin_path), 'log', '-1', '--format=%H%n%cI', 'HEAD'],
                 capture_output=True,
                 text=True,
                 timeout=10,
                 check=True
             )
-            sha = sha_result.stdout.strip()
-
-            branch_result = subprocess.run(
-                ['git', '-C', str(plugin_path), 'rev-parse', '--abbrev-ref', 'HEAD'],
-                capture_output=True,
-                text=True,
-                timeout=10,
-                check=True
-            )
-            branch = branch_result.stdout.strip()
-
-            if branch == 'HEAD':
-                branch = ''
-
-            # Get remote URL
-            remote_url_result = subprocess.run(
-                ['git', '-C', str(plugin_path), 'config', '--get', 'remote.origin.url'],
-                capture_output=True,
-                text=True,
-                timeout=10,
-                check=False
-            )
-            remote_url = remote_url_result.stdout.strip() if remote_url_result.returncode == 0 else None
-
-            # Get commit date in ISO format
-            date_result = subprocess.run(
-                ['git', '-C', str(plugin_path), 'log', '-1', '--format=%cI', 'HEAD'],
-                capture_output=True,
-                text=True,
-                timeout=10,
-                check=True
-            )
-            commit_date_iso = date_result.stdout.strip()
+            lines = log_result.stdout.strip().splitlines()
+            sha = lines[0] if lines else ''
+            commit_date_iso = lines[1] if len(lines) > 1 else ''
 
             result = {
                 'sha': sha,
                 'short_sha': sha[:7] if sha else '',
-                'branch': branch
+                'branch': branch,
             }
-            
-            # Add remote URL if available
+
             if remote_url:
                 result['remote_url'] = remote_url
 
-            # Add commit date if available
             if commit_date_iso:
                 result['date_iso'] = commit_date_iso
                 result['date'] = self._iso_to_date(commit_date_iso)

src/vegas_mode/coordinator.py

@@ -90,13 +90,11 @@ class VegasModeCoordinator:
         self._interrupt_check: Optional[Callable[[], bool]] = None
         self._interrupt_check_interval: int = 10  # Check every N frames
 
-        # Plugin update tick for keeping data fresh during Vegas mode
-        self._update_tick: Optional[Callable[[], Optional[List[str]]]] = None
-        self._update_tick_interval: float = 1.0  # Tick every 1 second
-        self._update_thread: Optional[threading.Thread] = None
-        self._update_results: Optional[List[str]] = None
-        self._update_results_lock = threading.Lock()
-        self._last_update_tick_time: float = 0.0
+        # Plugin update callback — fired from a background thread inside the loop
+        # so the main loop's _tick_plugin_updates() finds nothing due when Vegas
+        # returns, eliminating the inter-iteration frozen-frame gap.
+        self._update_callback: Optional[Callable[[], None]] = None
+        self._update_tick_running: bool = False
 
         # Config update tracking
         self._config_version = 0
@@ -139,6 +137,25 @@ class VegasModeCoordinator:
         """Check if Vegas mode is currently running."""
         return self._is_active
 
+    def set_sync_manager(self, sync_manager, follower_position: str = "left") -> None:
+        """
+        Attach a DisplaySyncManager so Vegas mode sends the follower's portion
+        of the ticker to the second display on every rendered frame.
+
+        Args:
+            sync_manager:       DisplaySyncManager instance, or None to disable sync
+            follower_position:  "left" (default) or "right" — physical position of
+                                the follower display relative to the leader
+        """
+        if self.render_pipeline:
+            # Don't expose a standalone (no-op) manager to the pipeline — treat it as None
+            if sync_manager is not None and hasattr(sync_manager, 'role'):
+                from src.common.sync_manager import SyncRole
+                if sync_manager.role == SyncRole.STANDALONE:
+                    sync_manager = None
+            self.render_pipeline.sync_manager = sync_manager
+            self.render_pipeline.sync_follower_left = (follower_position == "left")
+
     def set_live_priority_checker(self, checker: Callable[[], Optional[str]]) -> None:
         """
         Set the callback for checking live priority content.
@@ -166,24 +183,19 @@ class VegasModeCoordinator:
         self._interrupt_check = checker
         self._interrupt_check_interval = max(1, check_interval)
 
-    def set_update_tick(
-        self,
-        callback: Callable[[], Optional[List[str]]],
-        interval: float = 1.0
-    ) -> None:
+    def set_update_callback(self, callback: Callable[[], None]) -> None:
         """
-        Set the callback for periodic plugin update ticking during Vegas mode.
+        Set a callback for running plugin updates from inside the Vegas loop.
 
-        This keeps plugin data fresh while the Vegas render loop is running.
-        The callback should run scheduled plugin updates and return a list of
-        plugin IDs that were actually updated, or None/empty if no updates occurred.
+        Fired in a daemon background thread every ~4 s so plugin data stays
+        fresh without blocking the render loop.  The main loop's
+        _tick_plugin_updates() then finds all intervals already satisfied and
+        returns immediately, collapsing the inter-iteration gap to <1 ms.
 
         Args:
-            callback: Callable that returns list of updated plugin IDs or None
-            interval: Seconds between update tick calls (default 1.0)
+            callback: Callable with no arguments (typically _tick_plugin_updates)
         """
-        self._update_tick = callback
-        self._update_tick_interval = max(0.5, interval)
+        self._update_callback = callback
 
     def start(self) -> bool:
         """
@@ -237,9 +249,6 @@ class VegasModeCoordinator:
                 self.stats['total_runtime_seconds'] += time.time() - self._start_time
                 self._start_time = None
 
-        # Wait for in-flight background update before tearing down state
-        self._drain_update_thread()
-
         # Cleanup components
         self.render_pipeline.reset()
         self.stream_manager.reset()
@@ -335,83 +344,101 @@ class VegasModeCoordinator:
         last_fps_log_time = start_time
         fps_frame_count = 0
 
-        self._last_update_tick_time = start_time
-
         logger.info("Starting Vegas iteration for %.1fs", duration)
 
-        try:
-            while True:
-                # Check for STATIC mode plugin that should pause scroll
-                static_plugin = self._check_static_plugin_trigger()
-                if static_plugin:
-                    if not self._handle_static_pause(static_plugin):
-                        # Static pause was interrupted
+        while True:
+            # Check for STATIC mode plugin that should pause scroll
+            static_plugin = self._check_static_plugin_trigger()
+            if static_plugin:
+                if not self._handle_static_pause(static_plugin):
+                    # Static pause was interrupted
+                    return False
+                # After static pause, skip this segment and continue
+                self.stream_manager.get_next_segment()  # Consume the segment
+                continue
+
+            # Run frame
+            if not self.run_frame():
+                # Check why we stopped
+                with self._state_lock:
+                    if self._should_stop:
+                        return False
+                    if self._is_paused:
+                        # Paused for live priority - let caller handle
                         return False
-                    # After static pause, skip this segment and continue
-                    self.stream_manager.get_next_segment()  # Consume the segment
-                    continue
 
-                # Run frame
-                if not self.run_frame():
-                    # Check why we stopped
-                    with self._state_lock:
-                        if self._should_stop:
-                            return False
-                        if self._is_paused:
-                            # Paused for live priority - let caller handle
-                            return False
+            # Sleep for frame interval
+            time.sleep(frame_interval)
 
-                # Sleep for frame interval
-                time.sleep(frame_interval)
+            # Increment frame count and check for interrupt periodically
+            frame_count += 1
+            fps_frame_count += 1
 
-                # Increment frame count and check for interrupt periodically
-                frame_count += 1
-                fps_frame_count += 1
+            # Periodic FPS logging
+            current_time = time.time()
+            if current_time - last_fps_log_time >= fps_log_interval:
+                fps = fps_frame_count / (current_time - last_fps_log_time)
+                logger.info(
+                    "Vegas FPS: %.1f (target: %d, frames: %d)",
+                    fps, self.vegas_config.target_fps, fps_frame_count
+                )
+                last_fps_log_time = current_time
+                fps_frame_count = 0
 
-                # Periodic FPS logging
-                current_time = time.time()
-                if current_time - last_fps_log_time >= fps_log_interval:
-                    fps = fps_frame_count / (current_time - last_fps_log_time)
-                    logger.info(
-                        "Vegas FPS: %.1f (target: %d, frames: %d)",
-                        fps, self.vegas_config.target_fps, fps_frame_count
-                    )
-                    last_fps_log_time = current_time
-                    fps_frame_count = 0
+            if (self._interrupt_check and
+                    frame_count % self._interrupt_check_interval == 0):
+                try:
+                    if self._interrupt_check():
+                        logger.debug(
+                            "Vegas interrupted by callback after %d frames",
+                            frame_count
+                        )
+                        return False
+                except Exception:
+                    # Log but don't let interrupt check errors stop Vegas
+                    logger.exception("Interrupt check failed")
 
-                # Periodic plugin update tick to keep data fresh (non-blocking)
-                self._drive_background_updates()
-
-                if (self._interrupt_check and
-                        frame_count % self._interrupt_check_interval == 0):
+            # Fire plugin update tick in a background thread every ~4 s.
+            # Running it here (rather than only between iterations) means the
+            # main loop's _tick_plugin_updates() finds all intervals already
+            # satisfied on return, so the inter-iteration gap is <1 ms and the
+            # display never shows a frozen frame between iterations.
+            _UPDATE_TICK_FRAMES = max(1, int(self.vegas_config.target_fps * 4))  # every 4 s regardless of FPS
+            if (self._update_callback and
+                    frame_count % _UPDATE_TICK_FRAMES == 0 and
+                    not self._update_tick_running):
+                self._update_tick_running = True
+                def _run_tick(cb=self._update_callback):
                     try:
-                        if self._interrupt_check():
-                            logger.debug(
-                                "Vegas interrupted by callback after %d frames",
-                                frame_count
-                            )
-                            return False
-                    except Exception:
-                        # Log but don't let interrupt check errors stop Vegas
-                        logger.exception("Interrupt check failed")
+                        cb()
+                    finally:
+                        self._update_tick_running = False
+                threading.Thread(
+                    target=_run_tick, daemon=True, name="vegas-plugin-tick"
+                ).start()
 
-                # Check elapsed time
-                elapsed = time.time() - start_time
-                if elapsed >= duration:
-                    break
+            # Check elapsed time
+            elapsed = time.time() - start_time
+            if elapsed >= duration:
+                break
 
-                # Check for cycle completion
-                if self.render_pipeline.is_cycle_complete():
-                    break
+            # NOTE: do NOT break on is_cycle_complete() here.
+            # When multi-display sync is active, breaking exits run_iteration()
+            # which causes a 2-3s delay before start_new_cycle() is called on
+            # the next run_iteration(). During that gap the scroll advances into
+            # the pre-roll zone, then start_new_cycle() resets it — producing a
+            # second visible jump on the follower display ~2.5s after the first.
+            #
+            # Instead, run_frame() handles cycle completion directly (it calls
+            # start_new_cycle() in the very next frame, 8ms later), collapsing
+            # the two events into a single clean transition.
+            #
+            # Without sync, the iteration now runs to its full duration and may
+            # cycle content multiple times within one iteration — acceptable for
+            # a continuous ticker.
 
-            logger.info("Vegas iteration completed after %.1fs", time.time() - start_time)
-            return True
-
-        finally:
-            # Ensure background update thread finishes before the main loop
-            # resumes its own _tick_plugin_updates() calls, preventing concurrent
-            # run_scheduled_updates() execution.
-            self._drain_update_thread()
+        logger.info("Vegas iteration completed after %.1fs", time.time() - start_time)
+        return True
 
     def _check_live_priority(self) -> bool:
         """
@@ -500,71 +527,6 @@ class VegasModeCoordinator:
                 if self._pending_config is None:
                     self._pending_config_update = False
 
-    def _run_update_tick_background(self) -> None:
-        """Run the plugin update tick in a background thread.
-
-        Stores results for the render loop to pick up on its next iteration,
-        so the scroll never blocks on API calls.
-        """
-        try:
-            updated_plugins = self._update_tick()
-            if updated_plugins:
-                with self._update_results_lock:
-                    # Accumulate rather than replace to avoid losing notifications
-                    # if a previous result hasn't been picked up yet
-                    if self._update_results is None:
-                        self._update_results = updated_plugins
-                    else:
-                        self._update_results.extend(updated_plugins)
-        except Exception:
-            logger.exception("Background plugin update tick failed")
-
-    def _drain_update_thread(self, timeout: float = 2.0) -> None:
-        """Wait for any in-flight background update thread to finish.
-
-        Called when transitioning out of Vegas mode so the main-loop
-        ``_tick_plugin_updates`` call doesn't race with a still-running
-        background thread.
-        """
-        if self._update_thread is not None and self._update_thread.is_alive():
-            self._update_thread.join(timeout=timeout)
-            if self._update_thread.is_alive():
-                logger.warning(
-                    "Background update thread did not finish within %.1fs", timeout
-                )
-
-    def _drive_background_updates(self) -> None:
-        """Collect finished background update results and launch new ticks.
-
-        Safe to call from both the main render loop and the static-pause
-        wait loop so that plugin data stays fresh regardless of which
-        code path is active.
-        """
-        # 1. Collect results from a previously completed background update
-        with self._update_results_lock:
-            ready_results = self._update_results
-            self._update_results = None
-        if ready_results:
-            for pid in ready_results:
-                self.mark_plugin_updated(pid)
-
-        # 2. Kick off a new background update if interval elapsed and none running
-        current_time = time.time()
-        if (self._update_tick and
-                current_time - self._last_update_tick_time >= self._update_tick_interval):
-            thread_alive = (
-                self._update_thread is not None
-                and self._update_thread.is_alive()
-            )
-            if not thread_alive:
-                self._last_update_tick_time = current_time
-                self._update_thread = threading.Thread(
-                    target=self._run_update_tick_background,
-                    daemon=True,
-                    name="vegas-update-tick",
-                )
-                self._update_thread.start()
-
     def mark_plugin_updated(self, plugin_id: str) -> None:
         """
         Notify that a plugin's data has been updated.
@@ -683,8 +645,10 @@ class VegasModeCoordinator:
                     logger.info("Static pause interrupted by live priority")
                     return False
 
-                # Keep plugin data fresh during static pause
-                self._drive_background_updates()
+                # Yield immediately if multi-display follower mode becomes active
+                if self._interrupt_check and self._interrupt_check():
+                    logger.info("Static pause interrupted by sync follower mode")
+                    return False
 
                 # Sleep in small increments to remain responsive
                 time.sleep(0.1)

src/vegas_mode/plugin_adapter.py

@@ -329,50 +329,51 @@ class PluginAdapter:
             # Save display state to restore after
             original_image = self.display_manager.image.copy()
 
-            # Method 1: Try _create_scrolling_display (stocks pattern)
-            if hasattr(plugin, '_create_scrolling_display'):
-                logger.info(
-                    "[%s] Triggering via _create_scrolling_display()",
-                    plugin_id
-                )
-                try:
-                    plugin._create_scrolling_display()
-                    cached_image = getattr(scroll_helper, 'cached_image', None)
-                    if cached_image is not None and isinstance(cached_image, Image.Image):
-                        logger.info(
-                            "[%s] _create_scrolling_display() SUCCESS: %dx%d",
-                            plugin_id, cached_image.width, cached_image.height
-                        )
-                        return cached_image
-                except (AttributeError, TypeError, ValueError, OSError):
-                    logger.exception(
-                        "[%s] _create_scrolling_display() failed", plugin_id
-                    )
-
-            # Method 2: Try display(force_clear=True) which typically builds scroll content
-            if hasattr(plugin, 'display'):
-                logger.info(
-                    "[%s] Triggering via display(force_clear=True)",
-                    plugin_id
-                )
-                try:
-                    self.display_manager.clear()
-                    plugin.display(force_clear=True)
-                    cached_image = getattr(scroll_helper, 'cached_image', None)
-                    if cached_image is not None and isinstance(cached_image, Image.Image):
-                        logger.info(
-                            "[%s] display(force_clear=True) SUCCESS: %dx%d",
-                            plugin_id, cached_image.width, cached_image.height
-                        )
-                        return cached_image
+            with self.display_manager.capture_mode():
+                # Method 1: Try _create_scrolling_display (stocks pattern)
+                if hasattr(plugin, '_create_scrolling_display'):
                     logger.info(
-                        "[%s] display(force_clear=True) did not populate cached_image",
+                        "[%s] Triggering via _create_scrolling_display()",
                         plugin_id
                     )
-                except (AttributeError, TypeError, ValueError, OSError):
-                    logger.exception(
-                        "[%s] display(force_clear=True) failed", plugin_id
+                    try:
+                        plugin._create_scrolling_display()
+                        cached_image = getattr(scroll_helper, 'cached_image', None)
+                        if cached_image is not None and isinstance(cached_image, Image.Image):
+                            logger.info(
+                                "[%s] _create_scrolling_display() SUCCESS: %dx%d",
+                                plugin_id, cached_image.width, cached_image.height
+                            )
+                            return cached_image
+                    except (AttributeError, TypeError, ValueError, OSError):
+                        logger.exception(
+                            "[%s] _create_scrolling_display() failed", plugin_id
+                        )
+
+                # Method 2: Try display(force_clear=True) which typically builds scroll content
+                if hasattr(plugin, 'display'):
+                    logger.info(
+                        "[%s] Triggering via display(force_clear=True)",
+                        plugin_id
                     )
+                    try:
+                        self.display_manager.clear()
+                        plugin.display(force_clear=True)
+                        cached_image = getattr(scroll_helper, 'cached_image', None)
+                        if cached_image is not None and isinstance(cached_image, Image.Image):
+                            logger.info(
+                                "[%s] display(force_clear=True) SUCCESS: %dx%d",
+                                plugin_id, cached_image.width, cached_image.height
+                            )
+                            return cached_image
+                        logger.info(
+                            "[%s] display(force_clear=True) did not populate cached_image",
+                            plugin_id
+                        )
+                    except (AttributeError, TypeError, ValueError, OSError):
+                        logger.exception(
+                            "[%s] display(force_clear=True) failed", plugin_id
+                        )
 
             logger.info(
                 "[%s] Could not trigger scroll content generation",
@@ -408,10 +409,7 @@ class PluginAdapter:
             original_image = self.display_manager.image.copy()
             logger.info("[%s] Fallback: saved original display state", plugin_id)
 
-            # Lightweight in-memory data refresh before capturing.
-            # Full update() is intentionally skipped here — the background
-            # update tick in the Vegas coordinator handles periodic API
-            # refreshes so we don't block the content-fetch thread.
+            # Ensure plugin has fresh data before capturing
             has_update_data = hasattr(plugin, 'update_data')
             logger.info("[%s] Fallback: has update_data=%s", plugin_id, has_update_data)
             if has_update_data:
@@ -421,21 +419,24 @@ class PluginAdapter:
                 except (AttributeError, RuntimeError, OSError):
                     logger.exception("[%s] Fallback: update_data() failed", plugin_id)
 
-            # Clear and call plugin display
-            self.display_manager.clear()
-            logger.info("[%s] Fallback: display cleared, calling display()", plugin_id)
+            # Clear and call plugin display — use capture_mode to suppress hardware writes
+            # that plugins may trigger internally via update_display().
+            with self.display_manager.capture_mode():
+                self.display_manager.clear()
+                logger.info("[%s] Fallback: display cleared, calling display()", plugin_id)
 
-            # First try without force_clear (some plugins behave better this way)
-            try:
-                plugin.display()
-                logger.info("[%s] Fallback: display() called successfully", plugin_id)
-            except TypeError:
-                # Plugin may require force_clear argument
-                logger.info("[%s] Fallback: display() failed, trying with force_clear=True", plugin_id)
-                plugin.display(force_clear=True)
+                # First try without force_clear (some plugins behave better this way)
+                try:
+                    plugin.display()
+                    logger.info("[%s] Fallback: display() called successfully", plugin_id)
+                except TypeError:
+                    # Plugin may require force_clear argument
+                    logger.info("[%s] Fallback: display() failed, trying with force_clear=True", plugin_id)
+                    plugin.display(force_clear=True)
+
+                # Capture the result
+                captured = self.display_manager.image.copy()
 
-            # Capture the result
-            captured = self.display_manager.image.copy()
             logger.info(
                 "[%s] Fallback: captured frame %dx%d, mode=%s",
                 plugin_id, captured.width, captured.height, captured.mode
@@ -454,9 +455,10 @@ class PluginAdapter:
                     plugin_id
                 )
                 # Try once more with force_clear=True
-                self.display_manager.clear()
-                plugin.display(force_clear=True)
-                captured = self.display_manager.image.copy()
+                with self.display_manager.capture_mode():
+                    self.display_manager.clear()
+                    plugin.display(force_clear=True)
+                    captured = self.display_manager.image.copy()
 
                 is_blank, bright_ratio = self._is_blank_image(captured, return_ratio=True)
                 logger.info(
@@ -585,28 +587,6 @@ class PluginAdapter:
             else:
                 self._content_cache.clear()
 
-    def invalidate_plugin_scroll_cache(self, plugin: 'BasePlugin', plugin_id: str) -> None:
-        """
-        Clear a plugin's scroll_helper cache so Vegas re-fetches fresh visuals.
-
-        Uses scroll_helper.clear_cache() to reset all cached state (cached_image,
-        cached_array, total_scroll_width, scroll_position, etc.) — not just the
-        image.  Without this, plugins that use scroll_helper (stocks, news,
-        odds-ticker, etc.) would keep serving stale scroll images even after
-        their data refreshes.
-
-        Args:
-            plugin: Plugin instance
-            plugin_id: Plugin identifier
-        """
-        scroll_helper = getattr(plugin, 'scroll_helper', None)
-        if scroll_helper is None:
-            return
-
-        if getattr(scroll_helper, 'cached_image', None) is not None:
-            scroll_helper.clear_cache()
-            logger.debug("[%s] Cleared scroll_helper cache", plugin_id)
-
     def get_content_type(self, plugin: 'BasePlugin', plugin_id: str) -> str:
         """
         Get the type of content a plugin provides.

src/vegas_mode/render_pipeline.py

@@ -52,6 +52,10 @@ class RenderPipeline:
         self.config = config
         self.display_manager = display_manager
         self.stream_manager = stream_manager
+        self.sync_manager = None        # Optional DisplaySyncManager — set by coordinator
+        self.sync_follower_left = True  # True = follower is LEFT of leader (default)
+        self._sync_send_interval = 1.0 / 90  # raw bytes are cheap; 90fps > follower render rate
+        self._last_sync_send = 0.0
 
         # Display dimensions (handle both property and method access patterns)
         self.display_width = (
@@ -202,8 +206,26 @@ class RenderPipeline:
             # Update scroll position
             self.scroll_helper.update_scroll_position()
 
-            # Check if cycle is complete
-            if self.scroll_helper.is_scroll_complete():
+            # Determine if the cycle is done.
+            #
+            # scroll_helper considers a cycle complete only after
+            # total_distance_scrolled >= total_scroll_width + display_width.
+            # That extra display_width of travel causes a "wrap-around" phase
+            # where scroll_position resets to ~0 and the first plugin's content
+            # re-enters from the right — the user sees this 2-3 s of re-entry
+            # as "a plugin partially displaying before the next one starts."
+            #
+            # We end the cycle as soon as total_distance_scrolled reaches
+            # total_scroll_width (the wrap-around point), before any second-pass
+            # content becomes visible.  The scroll_helper's own is_scroll_complete()
+            # check is kept as a fallback for any edge-cases where that threshold
+            # is never hit.
+            at_wrap_point = (
+                not self._cycle_complete and
+                self.scroll_helper.total_distance_scrolled >= self.scroll_helper.total_scroll_width
+            )
+
+            if at_wrap_point or self.scroll_helper.is_scroll_complete():
                 if not self._cycle_complete:
                     self._cycle_complete = True
                     self.stats['scroll_cycles'] += 1
@@ -211,6 +233,17 @@ class RenderPipeline:
                         "Scroll cycle complete after %.1fs",
                         time.time() - self._cycle_start_time
                     )
+                    # Push blank immediately so the hardware never shows any
+                    # post-wrap content while the coordinator recomposes the
+                    # next cycle (~100 ms).
+                    try:
+                        from PIL import Image as _Image
+                        blank = _Image.new('RGB', (self.display_width, self.display_height))
+                        self.display_manager.image = blank
+                        self.display_manager.update_display()
+                    except Exception:
+                        logger.exception("Failed to write blank frame to display at cycle end")
+                return True  # Cycle done; coordinator starts new cycle next frame
 
             # Get visible portion
             visible_frame = self.scroll_helper.get_visible_portion()
@@ -221,6 +254,15 @@ class RenderPipeline:
             self.display_manager.image = visible_frame
             self.display_manager.update_display()
 
+            # Multi-display sync: send scroll position to follower.
+            # The follower renders from its own cached_array (kept identical to the
+            # leader's via TCP image transfer at each new_cycle) at scroll_x ± display_width.
+            if self.sync_manager:
+                now = time.time()
+                if now - self._last_sync_send >= self._sync_send_interval:
+                    self._last_sync_send = now
+                    self.sync_manager.send_scroll_x(self.scroll_helper.scroll_position)
+
             # Update scrolling state
             self.display_manager.set_scrolling_state(True)
 
@@ -260,33 +302,38 @@ class RenderPipeline:
         if self._cycle_complete:
             return True
 
+        # When multi-display sync is active, defer mid-cycle hot swaps until the
+        # cycle ends naturally. Hot swaps block the render loop for 15-30ms while
+        # the image is rebuilt, causing a freeze+jump that the follower perceives
+        # as a speed-up. Deferring to cycle boundaries keeps transitions clean.
+        # Staging buffer content is still pre-loaded; it just applies at cycle end.
+        if self.sync_manager is not None:
+            return False
+
         # Check if we need more content in the buffer
         buffer_status = self.stream_manager.get_buffer_status()
         if buffer_status['staging_count'] > 0:
             return True
 
-        # Trigger recompose when pending updates affect visible segments
-        if self.stream_manager.has_pending_updates_for_visible_segments():
-            return True
-
         return False
 
     def hot_swap_content(self) -> bool:
         """
         Hot-swap to new composed content.
 
-        Called when staging buffer has updated content or pending updates exist.
-        Preserves scroll position for mid-cycle updates to prevent visual jumps.
+        Called when staging buffer has updated content.
+        Swaps atomically to prevent visual glitches.
 
         Returns:
             True if swap occurred
         """
         try:
-            # Save scroll position for mid-cycle updates
-            saved_position = self.scroll_helper.scroll_position
-            saved_total_distance = self.scroll_helper.total_distance_scrolled
-            saved_total_width = max(1, self.scroll_helper.total_scroll_width)
-            was_mid_cycle = not self._cycle_complete
+            # Snapshot position before swap so we can reposition after.
+            # The new image has completely different content — if scroll_position
+            # is left unchanged it lands at an arbitrary mid-content point in the
+            # new image, causing a visible jump on both displays.
+            old_width = self.scroll_helper.total_scroll_width
+            old_pos = self.scroll_helper.scroll_position
 
             # Process any pending updates
             self.stream_manager.process_updates()
@@ -294,20 +341,24 @@ class RenderPipeline:
 
             # Recompose with updated content
             if self.compose_scroll_content():
+                # Map scroll position proportionally into the new image width so
+                # we resume at the same relative progress through the content.
+                # This keeps the visual tempo consistent and avoids the jump that
+                # occurred when old scroll_position landed arbitrarily in new image.
+                new_width = self.scroll_helper.total_scroll_width
+                if old_width > 0 and new_width > 0:
+                    ratio = (old_pos % old_width) / old_width
+                    self.scroll_helper.scroll_position = ratio * new_width
+                else:
+                    self.scroll_helper.scroll_position = 0.0
+
                 self.stats['hot_swaps'] += 1
-                # Restore scroll position for mid-cycle updates so the
-                # scroll continues from where it was instead of jumping to 0
-                if was_mid_cycle:
-                    new_total_width = max(1, self.scroll_helper.total_scroll_width)
-                    progress_ratio = min(saved_total_distance / saved_total_width, 0.999)
-                    self.scroll_helper.total_distance_scrolled = progress_ratio * new_total_width
-                    self.scroll_helper.scroll_position = min(
-                        saved_position,
-                        float(new_total_width - 1)
-                    )
-                    self.scroll_helper.scroll_complete = False
-                    self._cycle_complete = False
-                logger.debug("Hot-swap completed (mid_cycle_restore=%s)", was_mid_cycle)
+                logger.debug(
+                    "Hot-swap completed: scroll repositioned %.0f→%.0f (%.1f%% of new %dpx image)",
+                    old_pos, self.scroll_helper.scroll_position,
+                    (self.scroll_helper.scroll_position / new_width * 100) if new_width else 0,
+                    new_width,
+                )
                 return True
 
             return False
@@ -342,7 +393,29 @@ class RenderPipeline:
             return False
 
         # Compose new scroll content
-        return self.compose_scroll_content()
+        result = self.compose_scroll_content()
+
+        if result and self.sync_manager:
+            # When sync is active, start the leader at display_width instead of 0.
+            # This skips the initial black gap so the leader immediately shows content.
+            # The follower starts at position 0 (the gap) which looks like a clean
+            # blank transition rather than near-end content wrapping around.
+            self.scroll_helper.scroll_position = float(self.display_width)
+
+        if result and self.sync_manager:
+            # Signal follower that a new cycle started (triggers its own rebuild)
+            self.sync_manager.send_new_cycle()
+            # Push the actual scroll image over TCP so follower has identical pixels.
+            # Done in a background thread to not block the render loop (~15ms transfer).
+            if self.scroll_helper.cached_image is not None:
+                import threading as _t
+                _t.Thread(
+                    target=self.sync_manager.send_scroll_image,
+                    args=(self.scroll_helper.cached_image,),
+                    daemon=True, name="sync-image-push"
+                ).start()
+
+        return result
 
     def get_current_scroll_info(self) -> Dict[str, Any]:
         """Get current scroll state information."""

src/wifi_manager.py

@@ -60,12 +60,16 @@ def get_wifi_config_path():
 
 HOSTAPD_CONFIG_PATH = Path("/etc/hostapd/hostapd.conf")
 DNSMASQ_CONFIG_PATH = Path("/etc/dnsmasq.d/ledmatrix-captive.conf")
+# Drop-in config for NetworkManager's built-in dnsmasq (ipv4.method=shared).
+# Writing address=/#/<ap_ip> here causes NM to resolve every hostname to the AP,
+# triggering the OS captive-portal popup automatically on iOS/Android/Windows/macOS.
+NM_DNSMASQ_SHARED_DIR = Path("/etc/NetworkManager/dnsmasq-shared.d")
+NM_DNSMASQ_SHARED_CONF = NM_DNSMASQ_SHARED_DIR / "ledmatrix-captive.conf"
 HOSTAPD_SERVICE = "hostapd"
 DNSMASQ_SERVICE = "dnsmasq"
 
 # Default AP settings
 DEFAULT_AP_SSID = "LEDMatrix-Setup"
-DEFAULT_AP_PASSWORD = "ledmatrix123"
 DEFAULT_AP_CHANNEL = 7
 
 # LED status message file (for display_controller integration)
@@ -138,6 +142,11 @@ class WiFiManager:
         self._disconnected_checks = 0
         self._disconnected_checks_required = 3  # Require 3 consecutive disconnected checks (90 seconds at 30s interval)
 
+        # Timestamp set when AP mode is enabled; used for the idle-timeout check
+        self._ap_enabled_at: Optional[float] = None
+        # Which redirect backend was used (iptables/nftables/None); set per-instance
+        self._redirect_backend: Optional[str] = None
+
         logger.info(f"WiFi Manager initialized - nmcli: {self.has_nmcli}, iwlist: {self.has_iwlist}, "
                    f"hostapd: {self.has_hostapd}, dnsmasq: {self.has_dnsmasq}, "
                    f"interface: {self._wifi_interface}, trixie: {self._is_trixie}")
@@ -201,6 +210,24 @@ class WiFiManager:
         except (subprocess.TimeoutExpired, subprocess.SubprocessError, OSError):
             return False
 
+    def _find_command_path(self, command: str) -> Optional[str]:
+        """
+        Return the absolute path of a command, checking sbin locations that may not
+        be on PATH in restricted service environments.  Returns None if not found.
+        """
+        try:
+            result = subprocess.run(["which", command], capture_output=True,
+                                    text=True, timeout=2)
+            if result.returncode == 0 and result.stdout.strip():
+                return result.stdout.strip()
+        except (subprocess.TimeoutExpired, subprocess.SubprocessError, OSError):
+            pass
+        for path in [f"/usr/sbin/{command}", f"/sbin/{command}",
+                     f"/usr/local/sbin/{command}"]:
+            if os.path.isfile(path) and os.access(path, os.X_OK):
+                return path
+        return None
+
     def _discover_wifi_interface(self) -> str:
         """
         Discover the primary WiFi interface name dynamically.
@@ -303,7 +330,6 @@ class WiFiManager:
         else:
             self.config = {
                 "ap_ssid": DEFAULT_AP_SSID,
-                "ap_password": DEFAULT_AP_PASSWORD,
                 "ap_channel": DEFAULT_AP_CHANNEL,
                 "auto_enable_ap_mode": True,  # Default: auto-enable when no network (safe due to grace period)
                 "saved_networks": []
@@ -658,7 +684,286 @@ class WiFiManager:
             return False
         except (subprocess.TimeoutExpired, subprocess.SubprocessError, OSError):
             return False
-    
+
+    # ---------------------------------------------------------------------------
+    # Helpers
+    # ---------------------------------------------------------------------------
+
+    _IP_FORWARD_SAVE_PATH = Path("/tmp/ledmatrix_ip_forward_saved")
+
+    def _validate_ap_config(self) -> Tuple[str, int]:
+        """Return a sanitized (ssid, channel) pair from config, falling back to defaults."""
+        ssid = str(self.config.get("ap_ssid", DEFAULT_AP_SSID))
+        if not ssid or len(ssid) > 32 or not re.match(r'^[\x20-\x7E]+$', ssid):
+            logger.warning(f"AP SSID '{ssid}' is invalid, falling back to default")
+            ssid = DEFAULT_AP_SSID
+        try:
+            channel = int(self.config.get("ap_channel", DEFAULT_AP_CHANNEL))
+            if channel < 1 or channel > 14:
+                raise ValueError
+        except (TypeError, ValueError):
+            logger.warning("AP channel out of range, falling back to default")
+            channel = DEFAULT_AP_CHANNEL
+        return ssid, channel
+
+    def _setup_iptables_redirect(self) -> bool:
+        """
+        Add port 80 → 5000 redirect rules for the captive portal.
+
+        Tries iptables first, falls back to nftables (used by Debian Trixie).
+        When neither tool is available, logs a warning and returns True — the AP
+        still works and DNS spoofing still triggers the OS popup; users just land
+        on port 5000 directly rather than being redirected from port 80.
+
+        Only returns False when a tool was found but the rule addition itself failed.
+        """
+        try:
+            iptables = self._find_command_path("iptables")
+            nft = self._find_command_path("nft")
+
+            if not iptables and not nft:
+                logger.warning(
+                    "Neither iptables nor nft found; captive portal port-80 redirect unavailable. "
+                    "DNS spoofing will still trigger the OS popup but HTTP on port 80 won't reach Flask."
+                )
+                self._redirect_backend = None
+                return True  # AP works; redirect is best-effort
+
+            if iptables:
+                return self._setup_iptables_redirect_iptables(iptables)
+            else:
+                return self._setup_iptables_redirect_nftables(nft)
+
+        except Exception as e:
+            logger.warning(f"Could not set up port redirect: {e}")
+            try:
+                self._teardown_iptables_redirect()
+            except Exception as cleanup_e:
+                logger.warning(f"Cleanup after redirect exception also failed: {cleanup_e}")
+            return False
+
+    def _setup_iptables_redirect_iptables(self, iptables: str) -> bool:
+        """Set up port 80→5000 redirect using iptables."""
+        # Save ip_forward state before enabling
+        try:
+            current_fwd = Path("/proc/sys/net/ipv4/ip_forward").read_text().strip()
+        except OSError:
+            current_fwd = None
+        if current_fwd is not None:
+            try:
+                self._IP_FORWARD_SAVE_PATH.write_text(current_fwd)
+            except OSError:
+                current_fwd = None
+                logger.warning("Could not write ip_forward save file; state will not be restored")
+
+        if current_fwd != "1":
+            sysctl = self._find_command_path("sysctl")
+            sysctl_bin = sysctl if sysctl else "sysctl"
+            r = subprocess.run(["sudo", sysctl_bin, "-w", "net.ipv4.ip_forward=1"],
+                               capture_output=True, text=True, timeout=5)
+            if r.returncode != 0:
+                logger.error(f"Failed to enable ip_forward: {r.stderr.strip()}")
+                self._teardown_iptables_redirect()
+                return False
+
+        if subprocess.run(
+            ["sudo", iptables, "-t", "nat", "-C", "PREROUTING",
+             "-i", self._wifi_interface, "-p", "tcp", "--dport", "80",
+             "-j", "REDIRECT", "--to-port", "5000"],
+            capture_output=True, timeout=5
+        ).returncode != 0:
+            r = subprocess.run(
+                ["sudo", iptables, "-t", "nat", "-A", "PREROUTING",
+                 "-i", self._wifi_interface, "-p", "tcp", "--dport", "80",
+                 "-j", "REDIRECT", "--to-port", "5000"],
+                capture_output=True, text=True, timeout=5
+            )
+            if r.returncode != 0:
+                logger.error(f"Failed to add PREROUTING rule: {r.stderr.strip()}")
+                self._teardown_iptables_redirect()
+                return False
+
+        if subprocess.run(
+            ["sudo", iptables, "-C", "INPUT",
+             "-i", self._wifi_interface, "-p", "tcp", "--dport", "5000", "-j", "ACCEPT"],
+            capture_output=True, timeout=5
+        ).returncode != 0:
+            r = subprocess.run(
+                ["sudo", iptables, "-A", "INPUT",
+                 "-i", self._wifi_interface, "-p", "tcp", "--dport", "5000", "-j", "ACCEPT"],
+                capture_output=True, text=True, timeout=5
+            )
+            if r.returncode != 0:
+                logger.error(f"Failed to add INPUT rule: {r.stderr.strip()}")
+                self._teardown_iptables_redirect()
+                return False
+
+        self._redirect_backend = "iptables"
+        logger.info("iptables: port 80→5000 redirect rules added")
+        return True
+
+    def _setup_iptables_redirect_nftables(self, nft: str) -> bool:
+        """Set up port 80→5000 redirect using nftables (Debian Trixie / modern systems)."""
+        # NM's ipv4.method=shared already enables ip_forward; no sysctl needed.
+        cmds = [
+            ["sudo", nft, "add", "table", "ip", "ledmatrix"],
+            ["sudo", nft, "add", "chain", "ip", "ledmatrix", "prerouting",
+             "{", "type", "nat", "hook", "prerouting", "priority", "-100", ";", "}"],
+            ["sudo", nft, "add", "rule", "ip", "ledmatrix", "prerouting",
+             "iif", self._wifi_interface, "tcp", "dport", "80", "redirect", "to", ":5000"],
+        ]
+        for cmd in cmds:
+            r = subprocess.run(cmd, capture_output=True, text=True, timeout=5)
+            if r.returncode != 0:
+                # Table/chain may already exist — only fail on rule add
+                if "add rule" in " ".join(cmd):
+                    logger.error(f"Failed to add nftables redirect rule: {r.stderr.strip()}")
+                    self._teardown_iptables_redirect()
+                    return False
+                logger.debug(f"nft cmd non-zero (may already exist): {r.stderr.strip()}")
+
+        self._redirect_backend = "nftables"
+        logger.info("nftables: port 80→5000 redirect rule added")
+        return True
+
+    def _teardown_iptables_redirect(self) -> None:
+        """Remove the port 80→5000 redirect rules and restore ip_forward if saved."""
+        try:
+            backend = self._redirect_backend
+            self._redirect_backend = None
+
+            if backend == "iptables":
+                iptables = self._find_command_path("iptables")
+                if iptables:
+                    subprocess.run(
+                        ["sudo", iptables, "-t", "nat", "-D", "PREROUTING",
+                         "-i", self._wifi_interface, "-p", "tcp", "--dport", "80",
+                         "-j", "REDIRECT", "--to-port", "5000"],
+                        capture_output=True, timeout=5
+                    )
+                    subprocess.run(
+                        ["sudo", iptables, "-D", "INPUT",
+                         "-i", self._wifi_interface, "-p", "tcp", "--dport", "5000",
+                         "-j", "ACCEPT"],
+                        capture_output=True, timeout=5
+                    )
+                # Restore ip_forward only when we saved it
+                if self._IP_FORWARD_SAVE_PATH.exists():
+                    try:
+                        saved = self._IP_FORWARD_SAVE_PATH.read_text().strip()
+                        self._IP_FORWARD_SAVE_PATH.unlink(missing_ok=True)
+                        sysctl = self._find_command_path("sysctl")
+                        sysctl_bin = sysctl if sysctl else "sysctl"
+                        subprocess.run(["sudo", sysctl_bin, "-w", f"net.ipv4.ip_forward={saved}"],
+                                       capture_output=True, timeout=5)
+                        logger.info(f"ip_forward restored to {saved}")
+                    except OSError as e:
+                        logger.warning(f"Could not restore ip_forward: {e}")
+                else:
+                    logger.debug("ip_forward not modified by setup; leaving unchanged")
+
+            elif backend == "nftables":
+                nft = self._find_command_path("nft")
+                if nft:
+                    subprocess.run(
+                        ["sudo", nft, "delete", "table", "ip", "ledmatrix"],
+                        capture_output=True, timeout=5
+                    )
+                    logger.info("nftables ledmatrix table removed")
+
+            else:
+                # No redirect was set up (neither tool available); nothing to tear down
+                self._IP_FORWARD_SAVE_PATH.unlink(missing_ok=True)
+
+        except Exception as e:
+            logger.warning(f"Could not tear down port redirect: {e}")
+
+    def _write_nm_dnsmasq_captive_conf(self, ap_ip: str = "192.168.4.1") -> None:
+        """
+        Write the NM dnsmasq-shared.d drop-in that makes NM's built-in dnsmasq
+        resolve every hostname to the AP IP.  This triggers the OS captive-portal
+        popup automatically on iOS / Android / Windows / macOS as soon as the
+        device connects — no manual navigation required.
+
+        NetworkManager reads /etc/NetworkManager/dnsmasq-shared.d/*.conf when it
+        starts the dnsmasq instance for ipv4.method=shared connections.
+        """
+        try:
+            content = f"# LEDMatrix captive portal: resolve all hostnames to AP\naddress=/#/{ap_ip}\n"
+            with open("/tmp/ledmatrix-nm-dnsmasq.conf", "w") as f:
+                f.write(content)
+            subprocess.run(
+                ["sudo", "mkdir", "-p", str(NM_DNSMASQ_SHARED_DIR)],
+                capture_output=True, timeout=5
+            )
+            subprocess.run(
+                ["sudo", "cp", "/tmp/ledmatrix-nm-dnsmasq.conf", str(NM_DNSMASQ_SHARED_CONF)],
+                capture_output=True, timeout=5
+            )
+            logger.info(f"Wrote NM dnsmasq captive-portal config: {NM_DNSMASQ_SHARED_CONF}")
+        except Exception as e:
+            logger.warning(f"Could not write NM dnsmasq captive config: {e}")
+
+    def _remove_nm_dnsmasq_captive_conf(self) -> None:
+        """Remove the NM dnsmasq-shared.d drop-in written by _write_nm_dnsmasq_captive_conf."""
+        try:
+            subprocess.run(
+                ["sudo", "rm", "-f", str(NM_DNSMASQ_SHARED_CONF)],
+                capture_output=True, timeout=5
+            )
+            logger.info("Removed NM dnsmasq captive-portal config")
+        except Exception as e:
+            logger.warning(f"Could not remove NM dnsmasq captive config: {e}")
+
+    def _check_internet_connectivity(self, timeout: int = 5) -> bool:
+        """
+        Test actual internet reachability — not just nmcli association state.
+
+        A device can be 'connected' in nmcli (associated with an AP) while the
+        router has no WAN link.  This check catches that case so the daemon can
+        auto-enable AP mode even when nmcli reports a connection.
+
+        Returns True if at least one reachability method succeeds.
+        """
+        try:
+            r = subprocess.run(
+                ["ping", "-c", "1", "-W", str(timeout), "8.8.8.8"],
+                capture_output=True, timeout=timeout + 1
+            )
+            if r.returncode == 0:
+                logger.debug("Internet connectivity confirmed via ping 8.8.8.8")
+                return True
+        except (subprocess.SubprocessError, OSError):
+            pass
+        try:
+            import urllib.request as _ureq
+            _ureq.urlopen("http://connectivity-check.ubuntu.com/", timeout=timeout)
+            logger.debug("Internet connectivity confirmed via HTTP check")
+            return True
+        except OSError:
+            pass
+        logger.debug("Internet connectivity check failed (both ping and HTTP)")
+        return False
+
+    def check_internet_connectivity(self, timeout: int = 5) -> bool:
+        """Public wrapper around _check_internet_connectivity for use by the daemon."""
+        return self._check_internet_connectivity(timeout=timeout)
+
+    def _has_ap_clients(self) -> bool:
+        """
+        Return True if at least one client is associated with the AP.
+        Uses 'iw dev <iface> station dump' which works for both hostapd and
+        nmcli AP modes.
+        """
+        try:
+            result = subprocess.run(
+                ["iw", "dev", self._wifi_interface, "station", "dump"],
+                capture_output=True, text=True, timeout=5
+            )
+            return bool(result.stdout.strip())
+        except Exception:
+            return False
+
     def scan_networks(self, allow_cached: bool = True) -> Tuple[List[WiFiNetwork], bool]:
         """
         Scan for available WiFi networks.
@@ -1293,12 +1598,27 @@ class WiFiManager:
                 error_msg = result.stderr.strip() or result.stdout.strip()
                 logger.error(f"Failed to connect to {ssid}: {error_msg}")
                 self._show_led_message("Connection failed", duration=5)
+                if self._is_wrong_password_error(error_msg):
+                    return False, f"wrong_password: {error_msg}"
                 return False, error_msg
         except Exception as e:
             logger.error(f"Error connecting with nmcli: {e}")
             self._show_led_message("Connection error", duration=5)
             return False, str(e)
-    
+
+    @staticmethod
+    def _is_wrong_password_error(error_msg: str) -> bool:
+        """Return True when nmcli's error output indicates an authentication failure."""
+        indicators = [
+            "secrets were required",
+            "no secret agent",
+            "802-11-wireless-security.psk",
+            "authentication rejected",
+            "association rejected",
+        ]
+        lower = error_msg.lower()
+        return any(ind in lower for ind in indicators)
+
     def _connect_wpa_supplicant(self, ssid: str, password: str) -> Tuple[bool, str]:
         """Connect using wpa_supplicant (fallback)"""
         try:
@@ -1570,14 +1890,18 @@ class WiFiManager:
             if self.has_hostapd and self.has_dnsmasq:
                 result = self._enable_ap_mode_hostapd()
                 if result[0]:
+                    self._ap_enabled_at = time.time()
                     return result
-            
+
             # Fallback to nmcli hotspot (simpler, no captive portal)
             if self.has_nmcli:
                 logger.info("hostapd/dnsmasq failed or unavailable, trying nmcli hotspot fallback...")
                 self._show_led_message("Setup Mode", duration=5)
-                return self._enable_ap_mode_nmcli_hotspot()
-            
+                result = self._enable_ap_mode_nmcli_hotspot()
+                if result[0]:
+                    self._ap_enabled_at = time.time()
+                return result
+
             return False, "No WiFi tools available (nmcli, hostapd, or dnsmasq required)"
         except Exception as e:
             logger.error(f"Error in enable_ap_mode: {e}")
@@ -1649,63 +1973,21 @@ class WiFiManager:
                     subprocess.run(["sudo", "systemctl", "stop", HOSTAPD_SERVICE], timeout=5)
                     return False, f"Failed to start dnsmasq: {result.stderr}"
                 
-                # Set up iptables port forwarding: redirect port 80 to 5000
-                # This makes the captive portal work on standard HTTP port
-                try:
-                    # Check if iptables is available
-                    iptables_check = subprocess.run(
-                        ["which", "iptables"],
-                        capture_output=True,
-                        timeout=2
-                    )
-                    
-                    if iptables_check.returncode == 0:
-                        # Enable IP forwarding (needed for NAT)
-                        subprocess.run(
-                            ["sudo", "sysctl", "-w", "net.ipv4.ip_forward=1"],
-                            capture_output=True,
-                            timeout=5
-                        )
-                        
-                        # Add NAT rule to redirect port 80 to 5000 on WiFi interface
-                        # First check if rule already exists
-                        check_result = subprocess.run(
-                            ["sudo", "iptables", "-t", "nat", "-C", "PREROUTING", "-i", self._wifi_interface, "-p", "tcp", "--dport", "80", "-j", "REDIRECT", "--to-port", "5000"],
-                            capture_output=True,
-                            timeout=5
-                        )
+                # Set up iptables port forwarding (port 80 → 5000) and save ip_forward state
+                if not self._setup_iptables_redirect():
+                    logger.error("Captive-portal redirect setup failed; stopping AP services")
+                    subprocess.run(["sudo", "systemctl", "stop", HOSTAPD_SERVICE],
+                                   capture_output=True, timeout=10)
+                    subprocess.run(["sudo", "systemctl", "stop", DNSMASQ_SERVICE],
+                                   capture_output=True, timeout=10)
+                    return False, "AP started but captive-portal redirect setup failed"
 
-                        if check_result.returncode != 0:
-                            # Rule doesn't exist, add it
-                            subprocess.run(
-                                ["sudo", "iptables", "-t", "nat", "-A", "PREROUTING", "-i", self._wifi_interface, "-p", "tcp", "--dport", "80", "-j", "REDIRECT", "--to-port", "5000"],
-                                capture_output=True,
-                                timeout=5
-                            )
-                            logger.info("Added iptables rule to redirect port 80 to 5000")
-
-                        # Also allow incoming connections on port 80
-                        check_input = subprocess.run(
-                            ["sudo", "iptables", "-C", "INPUT", "-i", self._wifi_interface, "-p", "tcp", "--dport", "80", "-j", "ACCEPT"],
-                            capture_output=True,
-                            timeout=5
-                        )
-
-                        if check_input.returncode != 0:
-                            subprocess.run(
-                                ["sudo", "iptables", "-A", "INPUT", "-i", self._wifi_interface, "-p", "tcp", "--dport", "80", "-j", "ACCEPT"],
-                                capture_output=True,
-                                timeout=5
-                            )
-                    else:
-                        logger.debug("iptables not available, port forwarding not set up")
-                        logger.info("Note: Port 80 forwarding requires iptables. Users will need to access port 5000 directly.")
-                except Exception as e:
-                    logger.warning(f"Could not set up iptables port forwarding: {e}")
-                    # Continue anyway - port 5000 will still work
-                
                 logger.info("AP mode enabled successfully")
-                self._show_led_message("Setup Mode Active", duration=5)
+                # Use the validated SSID so the displayed name matches what hostapd broadcast
+                ap_ssid, _ = self._validate_ap_config()
+                self._show_led_message(
+                    f"WiFi Setup\n{ap_ssid}\nNo password\n192.168.4.1:5000", duration=10
+                )
                 return True, "AP mode enabled"
             except Exception as e:
                 logger.error(f"Error starting AP services: {e}")
@@ -1716,245 +1998,120 @@ class WiFiManager:
     
     def _enable_ap_mode_nmcli_hotspot(self) -> Tuple[bool, str]:
         """
-        Enable AP mode using nmcli hotspot.
+        Enable AP mode using nmcli as an open (passwordless) access point.
 
-        This method is optimized for both Bookworm and Trixie:
-        - Trixie: Uses Netplan, connections stored in /run/NetworkManager/system-connections
-        - Bookworm: Traditional NetworkManager, connections in /etc/NetworkManager/system-connections
+        Uses 'nmcli connection add type wifi 802-11-wireless.mode ap' instead of
+        'nmcli device wifi hotspot' because the hotspot subcommand always creates a
+        WPA2-protected network on Bookworm/Trixie and silently ignores attempts to
+        strip security after creation.
 
-        On Trixie, we also disable PMF (Protected Management Frames) which can cause
-        connection issues with certain WiFi adapters and clients.
+        Tested for both Bookworm and Trixie (Netplan-based NetworkManager).
         """
         try:
             # Stop any existing connection
             self.disconnect_from_network()
             time.sleep(1)
 
-            # Delete any existing hotspot connections (more thorough cleanup)
-            # First, list all connections to find any with the same SSID or hotspot-related ones
-            ap_ssid = self.config.get("ap_ssid", DEFAULT_AP_SSID)
-            result = subprocess.run(
-                ["nmcli", "-t", "-f", "NAME,TYPE,802-11-wireless.ssid", "connection", "show"],
-                capture_output=True,
-                text=True,
-                timeout=10
-            )
-            if result.returncode == 0:
-                for line in result.stdout.strip().split('\n'):
-                    if ':' in line:
-                        parts = line.split(':')
-                        if len(parts) >= 2:
-                            conn_name = parts[0].strip()
-                            conn_type = parts[1].strip().lower() if len(parts) > 1 else ""
-                            conn_ssid = parts[2].strip() if len(parts) > 2 else ""
+            ap_ssid, ap_channel = self._validate_ap_config()
 
-                            # Delete if:
-                            # 1. It's a hotspot type
-                            # 2. It has the same SSID as our AP
-                            # 3. It matches our known connection names
-                            should_delete = (
-                                'hotspot' in conn_type or
-                                conn_ssid == ap_ssid or
-                                'hotspot' in conn_name.lower() or
-                                conn_name in ["Hotspot", "LEDMatrix-Setup-AP", "TickerSetup-AP"]
-                            )
-
-                            if should_delete:
-                                logger.info(f"Deleting existing connection: {conn_name} (type: {conn_type}, SSID: {conn_ssid})")
-                                # First disconnect it if active
-                                subprocess.run(
-                                    ["nmcli", "connection", "down", conn_name],
-                                    capture_output=True,
-                                    timeout=5
-                                )
-                                # Then delete it
-                                subprocess.run(
-                                    ["nmcli", "connection", "delete", conn_name],
-                                    capture_output=True,
-                                    timeout=10
-                                )
-
-            # Also explicitly delete known connection names (in case they weren't caught above)
+            # Delete only the specific application-managed AP profiles by name.
+            # Never delete by SSID — that would destroy a user's saved home network.
             for conn_name in ["Hotspot", "LEDMatrix-Setup-AP", "TickerSetup-AP"]:
-                subprocess.run(
-                    ["nmcli", "connection", "down", conn_name],
-                    capture_output=True,
-                    timeout=5
-                )
-                subprocess.run(
-                    ["nmcli", "connection", "delete", conn_name],
-                    capture_output=True,
-                    timeout=10
-                )
+                subprocess.run(["nmcli", "connection", "down", conn_name],
+                                capture_output=True, timeout=5)
+                subprocess.run(["nmcli", "connection", "delete", conn_name],
+                                capture_output=True, timeout=10)
 
-            # Wait a moment for deletions to complete
             time.sleep(1)
 
-            # Get AP settings from config
-            ap_ssid = self.config.get("ap_ssid", DEFAULT_AP_SSID)
-            ap_channel = self.config.get("ap_channel", DEFAULT_AP_CHANNEL)
-
-            # Use nmcli hotspot command (simpler, works with Broadcom chips)
-            # Open network (no password) for easy setup access
-            logger.info(f"Creating open hotspot with nmcli: {ap_ssid} on {self._wifi_interface} (no password)")
-
-            # Note: Some NetworkManager versions add a default password to hotspots
-            # We'll create it and then immediately remove all security settings
+            # Create an open AP connection profile from scratch.
+            # Using 'connection add' instead of 'device wifi hotspot' because the
+            # hotspot subcommand always attaches a WPA2 PSK on Bookworm/Trixie and
+            # ignores post-creation security modifications.
+            logger.info(f"Creating open AP with nmcli connection add: {ap_ssid} on "
+                        f"{self._wifi_interface} (no password)")
             cmd = [
-                "nmcli", "device", "wifi", "hotspot",
-                "ifname", self._wifi_interface,
+                "nmcli", "connection", "add",
+                "type", "wifi",
                 "con-name", "LEDMatrix-Setup-AP",
+                "ifname", self._wifi_interface,
                 "ssid", ap_ssid,
-                "band", "bg",  # 2.4GHz for maximum compatibility
-                "channel", str(ap_channel),
-                # Don't pass password parameter - we'll remove security after creation
+                "802-11-wireless.mode", "ap",
+                "802-11-wireless.band", "bg",   # 2.4 GHz for maximum compatibility
+                "802-11-wireless.channel", str(ap_channel),
+                "ipv4.method", "shared",
+                "ipv4.addresses", "192.168.4.1/24",
+                # No 802-11-wireless-security section → open network
             ]
 
-            result = subprocess.run(
-                cmd,
-                capture_output=True,
-                text=True,
-                timeout=30
-            )
+            # PMF (Protected Management Frames) is only meaningful for WPA2/WPA3.
+            # An open AP has no security section, so adding 802-11-wireless-security.pmf
+            # would cause NM to require key-mgmt too, breaking the connection add on
+            # Trixie NM 1.52+. Leave PMF untouched — open APs have no frame protection.
 
-            if result.returncode == 0:
-                # Always explicitly remove all security settings to ensure open network
-                # NetworkManager sometimes adds default security even when not specified
-                logger.info("Ensuring hotspot is open (no password)...")
-                time.sleep(2)  # Give it a moment to create
+            result = subprocess.run(cmd, capture_output=True, text=True, timeout=30)
 
-                # Remove all possible security settings
-                security_settings = [
-                    ("802-11-wireless-security.key-mgmt", "none"),
-                    ("802-11-wireless-security.psk", ""),
-                    ("802-11-wireless-security.wep-key", ""),
-                    ("802-11-wireless-security.wep-key-type", ""),
-                    ("802-11-wireless-security.auth-alg", "open"),
-                ]
-
-                # On Trixie, also disable PMF (Protected Management Frames)
-                # This can cause connection issues with certain WiFi adapters and clients
-                if self._is_trixie:
-                    security_settings.append(("802-11-wireless-security.pmf", "disable"))
-                    logger.info("Trixie detected: disabling PMF for better client compatibility")
-
-                for setting, value in security_settings:
-                    result_modify = subprocess.run(
-                        ["nmcli", "connection", "modify", "LEDMatrix-Setup-AP", setting, str(value)],
-                        capture_output=True,
-                        text=True,
-                        timeout=5
-                    )
-                    if result_modify.returncode != 0:
-                        logger.debug(f"Could not set {setting} to {value}: {result_modify.stderr}")
-
-                # On Trixie, set static IP address for the hotspot (default is 10.42.0.1)
-                # We want 192.168.4.1 for consistency
-                subprocess.run(
-                    ["nmcli", "connection", "modify", "LEDMatrix-Setup-AP",
-                     "ipv4.addresses", "192.168.4.1/24",
-                     "ipv4.method", "shared"],
-                    capture_output=True,
-                    text=True,
-                    timeout=5
-                )
-                
-                # Verify it's open
-                verify_result = subprocess.run(
-                    ["nmcli", "-t", "-f", "802-11-wireless-security.key-mgmt,802-11-wireless-security.psk", "connection", "show", "LEDMatrix-Setup-AP"],
-                    capture_output=True,
-                    text=True,
-                    timeout=5
-                )
-                
-                if verify_result.returncode == 0:
-                    output = verify_result.stdout.strip()
-                    key_mgmt = ""
-                    psk = ""
-                    for line in output.split('\n'):
-                        if 'key-mgmt:' in line:
-                            key_mgmt = line.split(':', 1)[1].strip() if ':' in line else ""
-                        elif 'psk:' in line:
-                            psk = line.split(':', 1)[1].strip() if ':' in line else ""
-                    
-                    if key_mgmt != "none" or (psk and psk != ""):
-                        logger.warning(f"Hotspot still has security (key-mgmt={key_mgmt}, psk={'set' if psk else 'empty'}), deleting and recreating...")
-                        # Delete and recreate as last resort
-                        subprocess.run(
-                            ["nmcli", "connection", "down", "LEDMatrix-Setup-AP"],
-                            capture_output=True,
-                            timeout=5
-                        )
-                        subprocess.run(
-                            ["nmcli", "connection", "delete", "LEDMatrix-Setup-AP"],
-                            capture_output=True,
-                            timeout=5
-                        )
-                        time.sleep(1)
-                        # Recreate without any password parameters
-                        cmd_recreate = [
-                            "nmcli", "device", "wifi", "hotspot",
-                            "ifname", self._wifi_interface,
-                            "con-name", "LEDMatrix-Setup-AP",
-                            "ssid", ap_ssid,
-                            "band", "bg",
-                            "channel", str(ap_channel),
-                        ]
-                        subprocess.run(cmd_recreate, capture_output=True, timeout=30)
-                        # Set IP address for consistency
-                        subprocess.run(
-                            ["nmcli", "connection", "modify", "LEDMatrix-Setup-AP",
-                             "ipv4.addresses", "192.168.4.1/24",
-                             "ipv4.method", "shared"],
-                            capture_output=True,
-                            timeout=5
-                        )
-                        # Disable PMF on Trixie
-                        if self._is_trixie:
-                            subprocess.run(
-                                ["nmcli", "connection", "modify", "LEDMatrix-Setup-AP",
-                                 "802-11-wireless-security.pmf", "disable"],
-                                capture_output=True,
-                                timeout=5
-                            )
-                        logger.info("Recreated hotspot as open network")
-                    else:
-                        logger.info("Hotspot verified as open (no password)")
-                
-                # Restart the connection to apply all changes
-                subprocess.run(
-                    ["nmcli", "connection", "down", "LEDMatrix-Setup-AP"],
-                    capture_output=True,
-                    timeout=5
-                )
-                time.sleep(1)
-                subprocess.run(
-                    ["nmcli", "connection", "up", "LEDMatrix-Setup-AP"],
-                    capture_output=True,
-                    timeout=10
-                )
-                logger.info("Hotspot restarted with open network settings")
-                logger.info(f"AP mode started via nmcli hotspot: {ap_ssid}")
-                time.sleep(2)
-                
-                # Verify hotspot is running
-                status = self._get_ap_status_nmcli()
-                if status.get('active'):
-                    ip = status.get('ip', '192.168.4.1')
-                    logger.info(f"AP mode confirmed active at {ip}")
-                    self._show_led_message(f"Setup: {ip}", duration=5)
-                    return True, f"AP mode enabled (hotspot mode) - Access at {ip}:5000"
-                else:
-                    logger.error("AP mode started but not verified")
-                    return False, "AP mode started but verification failed"
-            else:
+            if result.returncode != 0:
                 error_msg = result.stderr.strip() or result.stdout.strip()
-                logger.error(f"Failed to start AP mode via nmcli: {error_msg}")
+                logger.error(f"Failed to create AP connection profile: {error_msg}")
                 self._show_led_message("AP mode failed", duration=5)
-                return False, f"Failed to start AP mode: {error_msg}"
-                
+                return False, f"Failed to create AP profile: {error_msg}"
+
+            # Write the NM dnsmasq-shared.d captive-portal config BEFORE bringing up
+            # the connection so NM's dnsmasq picks it up at start time.
+            # This causes every hostname DNS query from a connected device to resolve
+            # to 192.168.4.1, automatically triggering the OS captive-portal popup.
+            self._write_nm_dnsmasq_captive_conf()
+
+            logger.info("AP connection profile created, bringing it up...")
+            up_result = subprocess.run(
+                ["nmcli", "connection", "up", "LEDMatrix-Setup-AP"],
+                capture_output=True, text=True, timeout=20
+            )
+            if up_result.returncode != 0:
+                error_msg = up_result.stderr.strip() or up_result.stdout.strip()
+                logger.error(f"Failed to bring up AP connection: {error_msg}")
+                self._remove_nm_dnsmasq_captive_conf()
+                subprocess.run(["nmcli", "connection", "delete", "LEDMatrix-Setup-AP"],
+                               capture_output=True, timeout=10)
+                self._show_led_message("AP mode failed", duration=5)
+                return False, f"Failed to start AP: {error_msg}"
+
+            time.sleep(2)
+
+            # NM's ipv4.method=shared manages ip_forward automatically, so we only
+            # need to add the iptables port-redirect rules for the captive portal.
+            if not self._setup_iptables_redirect():
+                logger.error("Captive-portal redirect setup failed; rolling back AP profile")
+                self._remove_nm_dnsmasq_captive_conf()
+                subprocess.run(["nmcli", "connection", "down", "LEDMatrix-Setup-AP"],
+                               capture_output=True, timeout=10)
+                subprocess.run(["nmcli", "connection", "delete", "LEDMatrix-Setup-AP"],
+                               capture_output=True, timeout=10)
+                self._clear_led_message()
+                return False, "AP started but captive-portal redirect setup failed"
+
+            # Verify the AP is actually running
+            status = self._get_ap_status_nmcli()
+            if status.get('active'):
+                ip = status.get('ip', '192.168.4.1')
+                logger.info(f"AP mode confirmed active at {ip} (open network, no password)")
+                self._show_led_message(f"WiFi Setup\n{ap_ssid}\nNo password\n{ip}:5000", duration=10)
+                return True, f"AP mode enabled (open network) - Access at {ip}:5000"
+            else:
+                logger.error("AP mode started but not verified by status check — rolling back")
+                self._teardown_iptables_redirect()
+                self._remove_nm_dnsmasq_captive_conf()
+                subprocess.run(["nmcli", "connection", "down", "LEDMatrix-Setup-AP"],
+                               capture_output=True, timeout=10)
+                subprocess.run(["nmcli", "connection", "delete", "LEDMatrix-Setup-AP"],
+                               capture_output=True, timeout=10)
+                self._clear_led_message()
+                return False, "AP mode started but verification failed"
+
         except Exception as e:
-            logger.error(f"Error starting AP mode with nmcli hotspot: {e}")
+            logger.error(f"Error starting AP mode with nmcli: {e}")
+            self._remove_nm_dnsmasq_captive_conf()
             self._show_led_message("Setup mode error", duration=5)
             return False, str(e)
     
@@ -1976,7 +2133,12 @@ class WiFiManager:
 
             for line in result.stdout.strip().split('\n'):
                 parts = line.split(':')
-                if len(parts) >= 2 and 'hotspot' in parts[1].lower():
+                if len(parts) < 2:
+                    continue
+                conn_name = parts[0].strip()
+                conn_type = parts[1].strip().lower()
+                # Match our known AP profile name OR the legacy nmcli hotspot type
+                if conn_name == "LEDMatrix-Setup-AP" or 'hotspot' in conn_type:
                     # Get actual IP address (may be 192.168.4.1 or 10.42.0.1 depending on config)
                     ip = '192.168.4.1'
                     interface = parts[2] if len(parts) > 2 else self._wifi_interface
@@ -2072,45 +2234,9 @@ class WiFiManager:
                     except Exception as e:
                         logger.warning(f"Could not remove dnsmasq drop-in config: {e}")
                 
-                # Remove iptables port forwarding rules and disable IP forwarding (only for hostapd mode)
+                # Remove iptables redirect rules and restore ip_forward state (hostapd mode only)
                 if hostapd_active:
-                    try:
-                        # Check if iptables is available
-                        iptables_check = subprocess.run(
-                            ["which", "iptables"],
-                            capture_output=True,
-                            timeout=2
-                        )
-
-                        if iptables_check.returncode == 0:
-                            # Remove NAT redirect rule
-                            subprocess.run(
-                                ["sudo", "iptables", "-t", "nat", "-D", "PREROUTING", "-i", self._wifi_interface, "-p", "tcp", "--dport", "80", "-j", "REDIRECT", "--to-port", "5000"],
-                                capture_output=True,
-                                timeout=5
-                            )
-
-                            # Remove INPUT rule
-                            subprocess.run(
-                                ["sudo", "iptables", "-D", "INPUT", "-i", self._wifi_interface, "-p", "tcp", "--dport", "80", "-j", "ACCEPT"],
-                                capture_output=True,
-                                timeout=5
-                            )
-
-                            logger.info("Removed iptables port forwarding rules")
-                        else:
-                            logger.debug("iptables not available, skipping rule removal")
-
-                        # Disable IP forwarding (restore to default client mode)
-                        subprocess.run(
-                            ["sudo", "sysctl", "-w", "net.ipv4.ip_forward=0"],
-                            capture_output=True,
-                            timeout=5
-                        )
-                        logger.info("Disabled IP forwarding")
-                    except (subprocess.TimeoutExpired, subprocess.SubprocessError, OSError) as e:
-                        logger.warning(f"Could not remove iptables rules or disable forwarding: {e}")
-                        # Continue anyway
+                    self._teardown_iptables_redirect()
 
                     # Clean up WiFi interface IP configuration
                     subprocess.run(
@@ -2153,14 +2279,17 @@ class WiFiManager:
                         except Exception as e:
                             logger.error(f"Final WiFi radio unblock attempt failed: {e}")
                 else:
-                    # nmcli hotspot mode - restart not needed, just ensure WiFi radio is enabled
-                    logger.info("Skipping NetworkManager restart (nmcli hotspot mode, restart not needed)")
-                    # Still ensure WiFi radio is enabled (may have been disabled by nmcli operations)
-                    # Use retries for safety
+                    # nmcli AP mode — NM's ipv4.method=shared manages ip_forward automatically,
+                    # so we only need to remove the iptables redirect rules we added.
+                    logger.info("Skipping NetworkManager restart (nmcli AP mode, restart not needed)")
+                    self._teardown_iptables_redirect()
+                    self._remove_nm_dnsmasq_captive_conf()
+                    # Ensure WiFi radio is enabled after nmcli operations
                     wifi_enabled = self._ensure_wifi_radio_enabled(max_retries=3)
                     if not wifi_enabled:
-                        logger.warning("WiFi radio may be disabled after nmcli hotspot cleanup")
-                
+                        logger.warning("WiFi radio may be disabled after nmcli AP cleanup")
+
+                self._ap_enabled_at = None
                 logger.info("AP mode disabled successfully")
                 return True, "AP mode disabled"
             except Exception as e:
@@ -2175,9 +2304,11 @@ class WiFiManager:
         try:
             config_dir = HOSTAPD_CONFIG_PATH.parent
             config_dir.mkdir(parents=True, exist_ok=True)
-            
-            ap_ssid = self.config.get("ap_ssid", DEFAULT_AP_SSID)
-            ap_channel = self.config.get("ap_channel", DEFAULT_AP_CHANNEL)
+
+            # Use validated values — strips invalid chars and ensures channel is an int.
+            # Also strip newlines from SSID to prevent config-file injection.
+            ap_ssid, ap_channel = self._validate_ap_config()
+            ap_ssid = ap_ssid.replace('\n', '').replace('\r', '')
 
             # Open network configuration (no password) for easy setup access
             config_content = f"""interface={self._wifi_interface}
@@ -2305,22 +2436,21 @@ address=/detectportal.firefox.com/192.168.4.1
                         f"Ethernet={ethernet_connected}, AP_active={ap_active}, "
                         f"auto_enable={auto_enable}, disconnected_checks={self._disconnected_checks}")
             
-            # Determine if we should have AP mode active
-            # AP mode should only be auto-enabled if:
-            # - auto_enable_ap_mode is True AND
-            # - WiFi is NOT connected AND
-            # - Ethernet is NOT connected AND
-            # - We've had multiple consecutive disconnected checks (grace period)
+            # Determine if we should have AP mode active.
+            # AP-enable uses only the nmcli association state (fast, no network calls).
+            # This keeps the same reliable behaviour as before: momentary packet loss
+            # while on working WiFi does NOT trigger AP mode.  The internet-reachability
+            # check is performed separately in the daemon watchdog for NM recovery.
             is_disconnected = not status.connected and not ethernet_connected
-            
+
             if is_disconnected:
                 # Increment disconnected check counter
                 self._disconnected_checks += 1
                 logger.debug(f"Network disconnected (check {self._disconnected_checks}/{self._disconnected_checks_required})")
             else:
-                # Reset counter if we're connected
+                # Reset counter if we're associated
                 if self._disconnected_checks > 0:
-                    logger.debug(f"Network connected, resetting disconnected check counter")
+                    logger.debug("Network connected, resetting disconnected check counter")
                 self._disconnected_checks = 0
             
             # Only enable AP if we've had enough consecutive disconnected checks
@@ -2365,6 +2495,24 @@ address=/detectportal.firefox.com/192.168.4.1
                     # AP is active but auto_enable is disabled - this means it was manually enabled
                     # Don't disable it automatically, let it stay active
                     logger.debug("AP mode is active (manually enabled), keeping active")
+
+            # Idle-timeout check: disable AP if no client has connected within the window.
+            # Only applies when AP is active and we haven't just decided to enable/disable it.
+            if ap_active and self._ap_enabled_at is not None:
+                try:
+                    idle_timeout_min = max(1, min(1440, int(self.config.get("ap_idle_timeout_minutes", 15))))
+                except (TypeError, ValueError):
+                    idle_timeout_min = 15
+                elapsed = time.time() - self._ap_enabled_at
+                if elapsed > idle_timeout_min * 60 and not self._has_ap_clients():
+                    logger.info(
+                        f"AP idle timeout ({idle_timeout_min} min, no clients) — disabling AP"
+                    )
+                    success, message = self.disable_ap_mode()
+                    if success:
+                        return True
+                    else:
+                        logger.warning(f"Failed to disable AP on idle timeout: {message}")
             
             return False
         except Exception as e:

systemd/ledmatrix-wifi-monitor.service

@@ -7,7 +7,7 @@ Wants=network.target
 Type=simple
 User=root
 WorkingDirectory=__PROJECT_ROOT_DIR__
-ExecStart=/usr/bin/python3 /home/ledpi/LEDMatrix/scripts/utils/wifi_monitor_daemon.py --interval 30
+ExecStart=/usr/bin/python3 __PROJECT_ROOT_DIR__/scripts/utils/wifi_monitor_daemon.py --interval 30
 Restart=on-failure
 RestartSec=10
 StandardOutput=syslog

test/test_wifi_manager_ap.py

@@ -0,0 +1,334 @@
+"""
+Unit tests for WiFi AP mode — src.wifi_manager.
+
+Each test exercises logic that can be verified through the subprocess calls the
+manager emits, without requiring root access, hardware, or a running Pi.
+
+Scenarios covered:
+1. nmcli AP profile is created with no security parameters (open/passwordless).
+2. iptables PREROUTING and INPUT rules are added when the nmcli AP starts.
+3. iptables rules and ip_forward are reverted when the AP is torn down.
+4. LED matrix message includes the SSID, 'No password', and the setup URL.
+5. Known AP profile names are deleted before the new profile is created.
+"""
+from __future__ import annotations
+
+import json
+from pathlib import Path
+from unittest.mock import MagicMock, patch
+
+import pytest
+
+from src.wifi_manager import WiFiManager
+
+
+# ---------------------------------------------------------------------------
+# Helpers
+# ---------------------------------------------------------------------------
+
+def _ok(stdout: str = "", stderr: str = "") -> MagicMock:
+    r = MagicMock()
+    r.returncode = 0
+    r.stdout = stdout
+    r.stderr = stderr
+    return r
+
+
+def _fail(stdout: str = "", stderr: str = "error") -> MagicMock:
+    r = MagicMock()
+    r.returncode = 1
+    r.stdout = stdout
+    r.stderr = stderr
+    return r
+
+
+def _find_path_side_effect(name: str) -> str:
+    """Deterministic fake for _find_command_path."""
+    return {"iptables": "/usr/sbin/iptables", "sysctl": "/usr/sbin/sysctl"}.get(
+        name, f"/usr/bin/{name}"
+    )
+
+
+# ---------------------------------------------------------------------------
+# Fixtures
+# ---------------------------------------------------------------------------
+
+@pytest.fixture()
+def wifi_config(tmp_path: Path) -> Path:
+    """Minimal wifi_config.json in a temporary directory."""
+    cfg_dir = tmp_path / "config"
+    cfg_dir.mkdir()
+    cfg = {
+        "ap_ssid": "LEDMatrix-Setup",
+        "ap_channel": 7,
+        "auto_enable_ap_mode": True,
+        "saved_networks": [],
+    }
+    p = cfg_dir / "wifi_config.json"
+    p.write_text(json.dumps(cfg))
+    return p
+
+
+@pytest.fixture()
+def manager(wifi_config: Path, tmp_path: Path) -> WiFiManager:
+    """
+    WiFiManager with all system calls stubbed out during construction and the
+    ip_forward save file redirected to a per-test temporary path.
+    """
+    with patch("src.wifi_manager.subprocess.run", return_value=_ok(stdout="wlan0\n")), \
+         patch.object(WiFiManager, "_detect_trixie", return_value=False):
+        mgr = WiFiManager(config_path=wifi_config)
+
+    # Force clean, deterministic state regardless of what __init__ inferred
+    mgr._wifi_interface = "wlan0"
+    mgr.has_nmcli = True
+    mgr.has_hostapd = False
+    mgr.has_dnsmasq = False
+    mgr.has_iwlist = False
+    mgr._is_trixie = False
+    # Redirect the ip_forward save file to tmp so tests never share state
+    mgr._IP_FORWARD_SAVE_PATH = tmp_path / "ip_fwd_saved"
+    return mgr
+
+
+# ---------------------------------------------------------------------------
+# 1. AP profile is open (no password)
+# ---------------------------------------------------------------------------
+
+@pytest.mark.unit
+def test_nmcli_ap_profile_has_no_security_params(manager: WiFiManager) -> None:
+    """
+    The 'nmcli connection add' command must not include key-mgmt, psk, or any
+    WPA-related parameter.  On Bookworm/Trixie, NM creates a WPA2-protected
+    hotspot even when those values are set to 'none'/empty via a later
+    'connection modify', so the profile must be created without a security
+    section from the start.
+    """
+    captured: list[list[str]] = []
+
+    def _run(cmd, **kw):
+        captured.append(list(cmd))
+        return _ok()
+
+    with patch("src.wifi_manager.subprocess.run", side_effect=_run), \
+         patch.object(manager, "disconnect_from_network", return_value=(True, "ok")), \
+         patch.object(manager, "_setup_iptables_redirect", return_value=True), \
+         patch.object(manager, "_get_ap_status_nmcli",
+                      return_value={"active": True, "ip": "192.168.4.1"}), \
+         patch.object(manager, "_show_led_message"):
+
+        success, _ = manager._enable_ap_mode_nmcli_hotspot()
+
+    assert success, "AP enable should report success"
+
+    add_calls = [c for c in captured if "nmcli" in c and "connection" in c and "add" in c]
+    assert add_calls, "Expected at least one 'nmcli connection add' invocation"
+
+    add_str = " ".join(add_calls[0])
+    assert "key-mgmt" not in add_str, "AP profile must not set key-mgmt"
+    assert "psk" not in add_str, "AP profile must not include a PSK/password"
+    assert "wpa" not in add_str.lower(), "AP profile must not reference WPA"
+    assert "802-11-wireless.mode" in add_str, "AP profile must declare wireless mode"
+    # Verify the value for 802-11-wireless.mode is exactly "ap" — check the element
+    # that immediately follows the key in the command list, not a loose substring match.
+    cmd = add_calls[0]
+    try:
+        mode_idx = cmd.index("802-11-wireless.mode")
+        assert cmd[mode_idx + 1] == "ap", \
+            f"802-11-wireless.mode value must be exactly 'ap', got {cmd[mode_idx + 1]!r}"
+    except ValueError:
+        pytest.fail("802-11-wireless.mode not found as a list element in nmcli command")
+
+
+# ---------------------------------------------------------------------------
+# 2. iptables NAT rules are added when the AP starts
+# ---------------------------------------------------------------------------
+
+@pytest.mark.unit
+def test_iptables_nat_rules_added_on_ap_start(manager: WiFiManager) -> None:
+    """
+    _setup_iptables_redirect must add:
+    - a PREROUTING REDIRECT rule that maps incoming TCP port 80 to port 5000, and
+    - an INPUT ACCEPT rule for port 5000 (the post-redirect destination port,
+      NOT port 80 which never hits the INPUT chain after PREROUTING rewrites it).
+    """
+    captured: list[list[str]] = []
+
+    def _run(cmd, **kw):
+        captured.append(list(cmd))
+        # iptables -C (check) → rc=1 so the -A (add) branch executes
+        if "iptables" in " ".join(str(x) for x in cmd) and "-C" in cmd:
+            return _fail()
+        return _ok()
+
+    # Patch Path.read_text so /proc/sys/net/ipv4/ip_forward is readable on any OS
+    with patch("src.wifi_manager.subprocess.run", side_effect=_run), \
+         patch.object(manager, "_find_command_path", side_effect=_find_path_side_effect), \
+         patch("pathlib.Path.read_text", return_value="0\n"):
+
+        result = manager._setup_iptables_redirect()
+
+    assert result, "_setup_iptables_redirect must return True on success"
+
+    prerouting_adds = [c for c in captured if "iptables" in " ".join(c) and "-A" in c and "PREROUTING" in c]
+    assert prerouting_adds, "Expected 'iptables -A PREROUTING' invocation"
+    pr_str = " ".join(prerouting_adds[0])
+    assert "--dport" in pr_str and "80" in pr_str, "PREROUTING rule must match dport 80"
+    assert "5000" in pr_str, "PREROUTING rule must redirect to port 5000"
+    assert "REDIRECT" in pr_str, "PREROUTING rule must use REDIRECT target"
+
+    input_adds = [c for c in captured if "iptables" in " ".join(c) and "-A" in c and "INPUT" in c]
+    assert input_adds, "Expected 'iptables -A INPUT' invocation"
+    in_str = " ".join(input_adds[0])
+    assert "5000" in in_str, "INPUT rule must accept port 5000 (post-PREROUTING destination)"
+    assert "ACCEPT" in in_str, "INPUT rule must use ACCEPT target"
+
+    # Port 80 must NOT be used in the INPUT rule (it is already redirected by PREROUTING)
+    input_80 = [c for c in captured if "iptables" in " ".join(c) and "INPUT" in c and "--dport" in c and "80" in c]
+    assert not input_80, "INPUT rule must target port 5000, not port 80"
+
+
+# ---------------------------------------------------------------------------
+# 3a. iptables rules and ip_forward reverted on teardown
+# ---------------------------------------------------------------------------
+
+@pytest.mark.unit
+def test_iptables_rules_and_ip_forward_reverted_on_teardown(manager: WiFiManager) -> None:
+    """
+    _teardown_iptables_redirect must:
+    - remove the PREROUTING and INPUT iptables rules, and
+    - restore ip_forward to the exact value recorded in the save file.
+    """
+    original_fwd = "0"
+    manager._IP_FORWARD_SAVE_PATH.write_text(original_fwd)
+    # Teardown dispatches on the backend recorded during setup
+    manager._redirect_backend = "iptables"
+
+    captured: list[list[str]] = []
+
+    with patch("src.wifi_manager.subprocess.run",
+               side_effect=lambda cmd, **kw: (captured.append(list(cmd)) or _ok())), \
+         patch.object(manager, "_find_command_path", side_effect=_find_path_side_effect):
+
+        manager._teardown_iptables_redirect()
+
+    assert [c for c in captured if "iptables" in " ".join(c) and "-D" in c and "PREROUTING" in c], \
+        "Expected 'iptables -D PREROUTING' invocation"
+
+    assert [c for c in captured if "iptables" in " ".join(c) and "-D" in c and "INPUT" in c], \
+        "Expected 'iptables -D INPUT' invocation"
+
+    restore_calls = [
+        c for c in captured
+        if "sysctl" in " ".join(c) and f"ip_forward={original_fwd}" in " ".join(c)
+    ]
+    assert restore_calls, f"Expected sysctl to restore ip_forward to {original_fwd!r}"
+
+    assert not manager._IP_FORWARD_SAVE_PATH.exists(), \
+        "Save file must be removed after successful teardown"
+
+
+# ---------------------------------------------------------------------------
+# 3b. ip_forward untouched when no save file exists
+# ---------------------------------------------------------------------------
+
+@pytest.mark.unit
+def test_ip_forward_not_restored_when_save_file_absent(manager: WiFiManager) -> None:
+    """
+    When the save file is missing (setup never wrote it, e.g. because /proc was
+    unreadable or the write failed), teardown must NOT call sysctl so it does not
+    accidentally clobber ip_forward state owned by a VPN or NetworkManager.
+    """
+    assert not manager._IP_FORWARD_SAVE_PATH.exists()
+
+    captured: list[list[str]] = []
+
+    with patch("src.wifi_manager.subprocess.run",
+               side_effect=lambda cmd, **kw: (captured.append(list(cmd)) or _ok())), \
+         patch.object(manager, "_find_command_path", side_effect=_find_path_side_effect):
+
+        manager._teardown_iptables_redirect()
+
+    sysctl_calls = [
+        c for c in captured
+        if "sysctl" in " ".join(c) and "ip_forward" in " ".join(c)
+    ]
+    assert not sysctl_calls, \
+        "sysctl must not be called when no ip_forward save file exists"
+
+
+# ---------------------------------------------------------------------------
+# 4. LED message content
+# ---------------------------------------------------------------------------
+
+@pytest.mark.unit
+def test_led_message_shows_ssid_no_password_and_url(manager: WiFiManager) -> None:
+    """
+    When the nmcli AP activates, the LED message must include:
+    - the AP SSID ('LEDMatrix-Setup')
+    - the string 'No password'
+    - the AP IP address (192.168.4.1) and Flask port (5000)
+    """
+    led_messages: list[str] = []
+
+    with patch("src.wifi_manager.subprocess.run", return_value=_ok()), \
+         patch.object(manager, "disconnect_from_network", return_value=(True, "ok")), \
+         patch.object(manager, "_setup_iptables_redirect", return_value=True), \
+         patch.object(manager, "_get_ap_status_nmcli",
+                      return_value={"active": True, "ip": "192.168.4.1"}), \
+         patch.object(manager, "_show_led_message",
+                      side_effect=lambda msg, **kw: led_messages.append(msg)):
+
+        success, _ = manager._enable_ap_mode_nmcli_hotspot()
+
+    assert success, "AP enable should report success"
+    assert led_messages, "Expected at least one _show_led_message call"
+
+    combined = "\n".join(led_messages)
+    assert "No password" in combined, "LED message must say 'No password'"
+    assert "LEDMatrix-Setup" in combined, "LED message must include the AP SSID"
+    assert "192.168.4.1" in combined, "LED message must include the AP IP address"
+    assert "5000" in combined, "LED message must include the Flask port"
+
+
+# ---------------------------------------------------------------------------
+# 5. Stale AP profiles deleted before the new one is created
+# ---------------------------------------------------------------------------
+
+@pytest.mark.unit
+def test_existing_ap_profiles_deleted_before_new_profile_created(manager: WiFiManager) -> None:
+    """
+    Before 'nmcli connection add', the manager must issue
+    'nmcli connection down/delete' for every known AP profile name so stale
+    profiles (from a previous crash or partial setup) cannot block the new one.
+    """
+    captured: list[list[str]] = []
+
+    def _run(cmd, **kw):
+        captured.append(list(cmd))
+        return _ok()
+
+    with patch("src.wifi_manager.subprocess.run", side_effect=_run), \
+         patch.object(manager, "disconnect_from_network", return_value=(True, "ok")), \
+         patch.object(manager, "_setup_iptables_redirect", return_value=True), \
+         patch.object(manager, "_get_ap_status_nmcli",
+                      return_value={"active": True, "ip": "192.168.4.1"}), \
+         patch.object(manager, "_show_led_message"):
+
+        success, _ = manager._enable_ap_mode_nmcli_hotspot()
+
+    assert success
+
+    cmd_strs = [" ".join(c) for c in captured]
+
+    for profile in ("LEDMatrix-Setup-AP", "Hotspot", "TickerSetup-AP"):
+        assert any("connection delete" in s and profile in s for s in cmd_strs), \
+            f"Expected 'nmcli connection delete {profile}' before creating the new profile"
+
+    add_indices = [i for i, s in enumerate(cmd_strs) if "connection add" in s]
+    del_indices = [i for i, s in enumerate(cmd_strs) if "connection delete" in s]
+
+    assert add_indices, "Expected 'nmcli connection add' call"
+    assert del_indices, "Expected 'nmcli connection delete' calls"
+    assert max(del_indices) < min(add_indices), \
+        "All connection deletions must complete before the new profile is created"

web_interface/app.py

@@ -226,13 +226,24 @@ def serve_plugin_asset(plugin_id, filename):
                 'message': 'Internal server error'
             }), 500
 
+# Prime psutil CPU measurement once at startup so interval=None returns a real value
+try:
+    import psutil as _psutil_prime
+    _psutil_prime.cpu_percent(interval=None)
+except ImportError:
+    pass
+
 # Cached AP mode check — avoids creating a WiFiManager per request
 _ap_mode_cache = {'value': False, 'timestamp': 0}
-_AP_MODE_CACHE_TTL = 5  # seconds
+_AP_MODE_CACHE_TTL = 30  # seconds — AP mode is user-initiated; 30s is fine
+
+# Cached ledmatrix service status for SSE stats stream
+_ledmatrix_service_cache = {'active': False, 'timestamp': 0}
+_LEDMATRIX_SERVICE_CACHE_TTL = 15  # seconds
 
 def is_ap_mode_active():
     """
-    Check if access point mode is currently active (cached, 5s TTL).
+    Check if access point mode is currently active (cached, 30s TTL).
     Uses a direct systemctl check instead of instantiating WiFiManager.
     """
     now = time.time()
@@ -444,10 +455,11 @@ def system_status_generator():
             # Try to import psutil for system stats
             try:
                 import psutil
-                cpu_percent = round(psutil.cpu_percent(interval=1), 1)
+                # interval=None is non-blocking; primed at module startup above
+                cpu_percent = round(psutil.cpu_percent(interval=None), 1)
                 memory = psutil.virtual_memory()
                 memory_used_percent = round(memory.percent, 1)
-                
+
                 # Try to get CPU temperature (Raspberry Pi specific)
                 cpu_temp = 0
                 try:
@@ -455,20 +467,23 @@ def system_status_generator():
                         cpu_temp = round(float(f.read()) / 1000.0, 1)
                 except (OSError, ValueError):
                     pass
-                    
+
             except ImportError:
                 cpu_percent = 0
                 memory_used_percent = 0
                 cpu_temp = 0
-            
-            # Check if display service is running
-            service_active = False
-            try:
-                result = subprocess.run(['systemctl', 'is-active', 'ledmatrix'], 
-                                      capture_output=True, text=True, timeout=2)
-                service_active = result.stdout.strip() == 'active'
-            except (subprocess.SubprocessError, OSError):
-                pass
+
+            # Check if display service is running (cached to avoid per-client subprocess forks)
+            now = time.time()
+            if (now - _ledmatrix_service_cache['timestamp']) >= _LEDMATRIX_SERVICE_CACHE_TTL:
+                try:
+                    result = subprocess.run(['systemctl', 'is-active', 'ledmatrix'],
+                                            capture_output=True, text=True, timeout=2)
+                    _ledmatrix_service_cache['active'] = result.stdout.strip() == 'active'
+                except (subprocess.SubprocessError, OSError):
+                    pass
+                _ledmatrix_service_cache['timestamp'] = now
+            service_active = _ledmatrix_service_cache['active']
             
             status = {
                 'timestamp': time.time(),
@@ -546,7 +561,7 @@ def display_preview_generator():
         except Exception as e:
             yield {'error': str(e)}
         
-        time.sleep(0.5)  # Check 2 times per second (reduced frequency for better performance)
+        time.sleep(1.0)  # Check once per second — halves PIL encode overhead vs 0.5s
 
 # Logs generator for SSE
 def logs_generator():

web_interface/templates/v3/captive_setup.html

@@ -191,7 +191,10 @@ function doConnect() {
       // Poll for the new IP
       setTimeout(function() { checkNewIP(ssid); }, 3000);
     } else {
-      showMsg(data.message || 'Connection failed', 'err');
+      var msg = data.error_type === 'wrong_password'
+        ? 'Incorrect password — please try again'
+        : (data.message || 'Connection failed');
+      showMsg(msg, 'err');
       connecting = false;
       btn.disabled = false;
       btn.innerHTML = 'Connect';

web_interface/templates/v3/partials/display.html

@@ -49,9 +49,9 @@
                            name="chain_length"
                            value="{{ main_config.display.hardware.chain_length or 2 }}"
                            min="1"
-                           max="32"
+                           max="8"
                            class="form-control">
-                    <p class="mt-1 text-sm text-gray-600">Number of LED panels chained together (e.g. 2 for 128×32, 5 for 320×32)</p>
+                    <p class="mt-1 text-sm text-gray-600">Number of LED panels chained together</p>
                 </div>
 
                 <div class="form-group">
@@ -380,12 +380,74 @@
                         <!-- Plugin order list will be populated by JavaScript -->
                         <p class="text-sm text-gray-500 italic">Loading plugins...</p>
                     </div>
-                    <input type="hidden" id="vegas_plugin_order_value" name="vegas_plugin_order" value="{{ main_config.display.get('vegas_scroll', {}).get('plugin_order', [])|tojson }}">
-                    <input type="hidden" id="vegas_excluded_plugins_value" name="vegas_excluded_plugins" value="{{ main_config.display.get('vegas_scroll', {}).get('excluded_plugins', [])|tojson }}">
+                    <input type="hidden" id="vegas_plugin_order_value" name="vegas_plugin_order" value='{{ main_config.display.get("vegas_scroll", {}).get("plugin_order", [])|tojson }}'>
+                    <input type="hidden" id="vegas_excluded_plugins_value" name="vegas_excluded_plugins" value='{{ main_config.display.get("vegas_scroll", {}).get("excluded_plugins", [])|tojson }}'>
                 </div>
             </div>
         </div>
 
+        <!-- Multi-Display Sync Settings -->
+        <div class="bg-gray-50 rounded-lg p-4 mt-6">
+            <div class="flex items-center justify-between mb-4">
+                <div>
+                    <h3 class="text-md font-medium text-gray-900">
+                        <i class="fas fa-clone mr-2"></i>Multi-Display Sync
+                    </h3>
+                    <p class="mt-1 text-sm text-gray-600">
+                        Extend scrolling content across two LED matrix display units over WiFi.
+                        Both displays must have identical rows and cols. Chain length may differ.
+                    </p>
+                </div>
+            </div>
+
+            <div class="grid grid-cols-1 md:grid-cols-2 gap-4">
+                <div class="form-group">
+                    <label for="sync_role" class="block text-sm font-medium text-gray-700">Role</label>
+                    <select id="sync_role" name="sync_role" class="form-control" onchange="updateSyncUI()">
+                        <option value="standalone" {% if main_config.get('sync', {}).get('role', 'standalone') == 'standalone' %}selected{% endif %}>Standalone (disabled)</option>
+                        <option value="leader"     {% if main_config.get('sync', {}).get('role', 'standalone') == 'leader'     %}selected{% endif %}>Leader (drives scroll)</option>
+                        <option value="follower"   {% if main_config.get('sync', {}).get('role', 'standalone') == 'follower'   %}selected{% endif %}>Follower (receives frames)</option>
+                    </select>
+                    <p class="mt-1 text-sm text-gray-600">Set Leader on one Pi, Follower on the other. Restart required after changing.</p>
+                </div>
+
+                <div class="form-group">
+                    <label for="sync_port" class="block text-sm font-medium text-gray-700">UDP Port</label>
+                    <input type="number"
+                           id="sync_port"
+                           name="sync_port"
+                           value="{{ main_config.get('sync', {}).get('port', 5765) }}"
+                           min="1024"
+                           max="65535"
+                           class="form-control">
+                    <p class="mt-1 text-sm text-gray-600">
+                        Must match on both Pis. If ufw is active:
+                        <code class="text-xs bg-gray-200 px-1 rounded">sudo ufw allow {{ main_config.get('sync', {}).get('port', 5765) }}/udp</code>
+                    </p>
+                </div>
+
+                <div class="form-group" id="sync_position_group" style="display:none">
+                    <label for="sync_follower_position" class="block text-sm font-medium text-gray-700">Position</label>
+                    <select id="sync_follower_position" name="sync_follower_position" class="form-control">
+                        <option value="left"  {% if main_config.get('sync', {}).get('follower_position', 'left') == 'left'  %}selected{% endif %}>Left of leader</option>
+                        <option value="right" {% if main_config.get('sync', {}).get('follower_position', 'left') == 'right' %}selected{% endif %}>Right of leader</option>
+                    </select>
+                    <p class="mt-1 text-sm text-gray-600">Which side of the leader display this unit sits on.</p>
+                </div>
+            </div>
+
+            <!-- Live status indicator (populated by JS) -->
+            <div id="sync_status_bar" class="mt-4 hidden">
+                <div id="sync_status_content" class="flex items-start space-x-2 p-3 rounded-lg border text-sm"></div>
+            </div>
+
+            <!-- Incompatibility detail (shown when error) -->
+            <div id="sync_error_detail" class="mt-2 hidden">
+                <p class="text-xs text-yellow-700 bg-yellow-50 border border-yellow-200 rounded p-2" id="sync_error_text"></p>
+                <p class="text-xs text-gray-500 mt-1">rows and cols must match between displays. chain_length may differ.</p>
+            </div>
+        </div>
+
         <!-- Submit Button -->
         <div class="flex justify-end">
             <button type="submit"
@@ -643,4 +705,111 @@ if (typeof window.fixInvalidNumberInputs !== 'function') {
         initPluginOrderList();
     }
 })();
+
+// Multi-Display Sync UI
+(function() {
+    function updateSyncUI() {
+        const role = document.getElementById('sync_role').value;
+        const bar = document.getElementById('sync_status_bar');
+        const posGroup = document.getElementById('sync_position_group');
+        if (role === 'standalone') {
+            bar.classList.add('hidden');
+            document.getElementById('sync_error_detail').classList.add('hidden');
+            posGroup.style.display = 'none';
+        } else {
+            bar.classList.remove('hidden');
+            posGroup.style.display = role === 'follower' ? '' : 'none';
+            pollSyncStatus();
+        }
+    }
+    window.updateSyncUI = updateSyncUI;
+
+    function pollSyncStatus() {
+        const role = document.getElementById('sync_role') && document.getElementById('sync_role').value;
+        if (!role || role === 'standalone') return;
+
+        fetch('/api/v3/sync/status')
+            .then(r => r.json())
+            .then(resp => {
+                const d = resp.data || {};
+                renderSyncStatus(d);
+            })
+            .catch(() => {
+                renderSyncStatus({state: 'unknown'});
+            });
+    }
+
+    function renderSyncStatus(d) {
+        const content = document.getElementById('sync_status_content');
+        const errorDetail = document.getElementById('sync_error_detail');
+        const errorText = document.getElementById('sync_error_text');
+        if (!content) return;
+
+        const state = d.state || 'unknown';
+        const role  = d.role  || 'unknown';
+
+        let icon, colorClass, text;
+
+        if (state === 'connected' || state === 'follower') {
+            icon = '●';
+            colorClass = 'bg-green-50 border-green-200 text-green-800';
+            const peer = d.peer_ip || d.leader_ip || 'peer';
+            text = role === 'leader'
+                ? `Follower connected — ${peer} (chain ${d.peer_chain || '?'})`
+                : `Receiving from leader — ${peer}`;
+            errorDetail.classList.add('hidden');
+
+        } else if (state === 'incompatible') {
+            icon = '⚠';
+            colorClass = 'bg-yellow-50 border-yellow-200 text-yellow-800';
+            text = `Follower connected but incompatible panels`;
+            if (d.error) {
+                errorText.textContent = d.error;
+                errorDetail.classList.remove('hidden');
+            }
+
+        } else if (state === 'no_peer' || state === 'standalone') {
+            icon = '○';
+            colorClass = 'bg-gray-50 border-gray-200 text-gray-600';
+            text = role === 'leader' ? 'No follower detected' : 'Searching for leader…';
+            errorDetail.classList.add('hidden');
+
+        } else if (state === 'starting') {
+            icon = '○';
+            colorClass = 'bg-gray-50 border-gray-200 text-gray-500';
+            text = 'Display process starting…';
+            errorDetail.classList.add('hidden');
+
+        } else {
+            icon = '✕';
+            colorClass = 'bg-red-50 border-red-200 text-red-700';
+            text = 'Sync status unavailable';
+            errorDetail.classList.add('hidden');
+        }
+
+        content.className = `flex items-start space-x-2 p-3 rounded-lg border text-sm ${colorClass}`;
+        content.textContent = '';
+        const iconSpan = document.createElement('span');
+        iconSpan.className = 'font-bold text-lg leading-none';
+        iconSpan.textContent = icon;
+        const textSpan = document.createElement('span');
+        textSpan.textContent = text;
+        content.appendChild(iconSpan);
+        content.appendChild(textSpan);
+    }
+
+    // Initial UI state and polling — guard against duplicate intervals on re-run
+    function startSyncPolling() {
+        updateSyncUI();
+        if (!window.syncStatusInterval) {
+            window.syncStatusInterval = setInterval(pollSyncStatus, 5000);
+        }
+    }
+
+    if (document.readyState === 'loading') {
+        document.addEventListener('DOMContentLoaded', startSyncPolling);
+    } else {
+        startSyncPolling();
+    }
+})();
 </script>

web_interface/templates/v3/partials/plugins.html

@@ -31,7 +31,52 @@
             </div>
             <div id="installed-plugins-content" class="block">
                 <div id="installed-plugins-grid" class="grid grid-cols-1 md:grid-cols-2 lg:grid-cols-3 xl:grid-cols-4 2xl:grid-cols-5 gap-6">
-                    <!-- Plugins will be loaded here -->
+                    <!-- Skeleton cards shown while installed plugins load -->
+                    <div class="installed-skeleton plugin-card animate-pulse">
+                        <div class="flex items-start justify-between mb-4">
+                            <div class="flex-1 min-w-0 space-y-2">
+                                <div class="h-4 bg-gray-200 rounded w-3/4"></div>
+                                <div class="h-3 bg-gray-200 rounded w-1/2"></div>
+                                <div class="h-3 bg-gray-200 rounded w-2/3"></div>
+                            </div>
+                            <div class="h-6 w-10 bg-gray-200 rounded-full ml-4 flex-shrink-0"></div>
+                        </div>
+                        <div class="space-y-2 mb-4">
+                            <div class="h-3 bg-gray-200 rounded w-full"></div>
+                            <div class="h-3 bg-gray-200 rounded w-5/6"></div>
+                        </div>
+                        <div class="h-8 bg-gray-200 rounded w-full mt-auto"></div>
+                    </div>
+                    <div class="installed-skeleton plugin-card animate-pulse hidden md:block">
+                        <div class="flex items-start justify-between mb-4">
+                            <div class="flex-1 min-w-0 space-y-2">
+                                <div class="h-4 bg-gray-200 rounded w-2/3"></div>
+                                <div class="h-3 bg-gray-200 rounded w-1/3"></div>
+                                <div class="h-3 bg-gray-200 rounded w-1/2"></div>
+                            </div>
+                            <div class="h-6 w-10 bg-gray-200 rounded-full ml-4 flex-shrink-0"></div>
+                        </div>
+                        <div class="space-y-2 mb-4">
+                            <div class="h-3 bg-gray-200 rounded w-full"></div>
+                            <div class="h-3 bg-gray-200 rounded w-4/5"></div>
+                        </div>
+                        <div class="h-8 bg-gray-200 rounded w-full mt-auto"></div>
+                    </div>
+                    <div class="installed-skeleton plugin-card animate-pulse hidden lg:block">
+                        <div class="flex items-start justify-between mb-4">
+                            <div class="flex-1 min-w-0 space-y-2">
+                                <div class="h-4 bg-gray-200 rounded w-4/5"></div>
+                                <div class="h-3 bg-gray-200 rounded w-2/5"></div>
+                                <div class="h-3 bg-gray-200 rounded w-3/5"></div>
+                            </div>
+                            <div class="h-6 w-10 bg-gray-200 rounded-full ml-4 flex-shrink-0"></div>
+                        </div>
+                        <div class="space-y-2 mb-4">
+                            <div class="h-3 bg-gray-200 rounded w-full"></div>
+                            <div class="h-3 bg-gray-200 rounded w-3/4"></div>
+                        </div>
+                        <div class="h-8 bg-gray-200 rounded w-full mt-auto"></div>
+                    </div>
                 </div>
             </div>
         </div>
@@ -203,14 +248,30 @@
             </div>
 
             <div id="plugin-store-grid" class="grid grid-cols-1 md:grid-cols-2 lg:grid-cols-3 xl:grid-cols-4 2xl:grid-cols-5 gap-6">
-                <!-- Loading skeleton -->
+                <!-- Loading skeleton — hidden by showStoreLoading(false) when data arrives -->
                 <div class="store-loading col-span-full">
-                    <div class="grid grid-cols-1 md:grid-cols-2 lg:grid-cols-3 xl:grid-cols-4 2xl:grid-cols-5 gap-4">
-                        <div class="bg-gray-200 rounded-lg p-4 h-48 animate-pulse"></div>
-                        <div class="bg-gray-200 rounded-lg p-4 h-48 animate-pulse"></div>
-                        <div class="bg-gray-200 rounded-lg p-4 h-48 animate-pulse"></div>
-                        <div class="bg-gray-200 rounded-lg p-4 h-48 animate-pulse"></div>
-                        <div class="bg-gray-200 rounded-lg p-4 h-48 animate-pulse"></div>
+                    <div class="grid grid-cols-1 md:grid-cols-2 lg:grid-cols-3 xl:grid-cols-4 2xl:grid-cols-5 gap-6">
+                        {% for _ in range(10) %}
+                        <div class="plugin-card animate-pulse">
+                            <div class="flex items-start justify-between mb-4">
+                                <div class="flex-1 min-w-0 space-y-2">
+                                    <div class="h-4 bg-gray-200 rounded w-3/4"></div>
+                                    <div class="h-3 bg-gray-200 rounded w-1/2"></div>
+                                </div>
+                                <div class="h-5 w-14 bg-gray-200 rounded-full ml-3 flex-shrink-0"></div>
+                            </div>
+                            <div class="space-y-2 mb-4">
+                                <div class="h-3 bg-gray-200 rounded w-full"></div>
+                                <div class="h-3 bg-gray-200 rounded w-5/6"></div>
+                                <div class="h-3 bg-gray-200 rounded w-4/6"></div>
+                            </div>
+                            <div class="flex gap-2 mt-auto">
+                                <div class="h-3 bg-gray-200 rounded w-12"></div>
+                                <div class="h-3 bg-gray-200 rounded w-16"></div>
+                            </div>
+                            <div class="h-8 bg-gray-200 rounded w-full mt-3"></div>
+                        </div>
+                        {% endfor %}
                     </div>
                 </div>
             </div>