fix(devices): SP110E vendor handshake + Windows/bleak robustness

SP110E peripherals silently tear down the GATT link ~1s after connect
unless a two-write vendor handshake (01 00 → FFE2, 01 B7 E3 D5 → FFE1)
arrives immediately. Without it the first real write hangs 30s then
reconnect-loops forever. Adds optional BLEProtocol.init_writes executed
on connect, plumbs a per-write char_uuid through both transports, and
fixes the SP110E color/power frames from an incorrect 5 bytes to the
documented 4 bytes.

Windows/WinRT robustness:
- asyncio.wait_for hangs on bleak because WinRT IAsyncOperations refuse
  to cancel. _bounded_await() uses asyncio.wait() instead so timeouts
  actually return control even when the inner task is uncancellable.
- BleakClient connect by raw MAC string times out when WinRT guesses
  address type wrong; switched to pre-scanning with BleakScanner and
  passing the resolved BLEDevice, which carries the address type.
- Target-start fetch timeout bumped to 30s with retry disabled so the
  UI doesn't abort during the BLE pre-scan + connect + handshake path.

UI:
- Settings modal exposes Protocol Family (IconSelect grid, shared with
  add-device via parameterized ensureBleFamilyIconSelect) so users can
  fix a wrong family pick without recreating the device. Govee AES key
  row toggles on/off with family selection.

Also turns LAN auth back on in default_config.yaml, logs start_processing
requests on entry for easier diagnosis, and captures the full debug trail
in docs/BLE_LED_CONTROLLERS.md for future BLE work.

Refs the mbullington SP110E protocol gist for the handshake bytes.

docs/BLE_LED_CONTROLLERS.md

@@ -0,0 +1,109 @@
+# BLE LED Controllers — Investigation & Implementation Notes
+
+Reference for anyone touching the BLE device provider (`server/src/ledgrab/core/devices/ble_*`). Captures the protocol quirks, Windows/bleak traps, and hardware lockdown we hit while bringing up SP110E / Triones / Zengge / Govee support.
+
+## Architecture
+
+```
+BLEDeviceProvider  →  BLEClient  →  BLETransport (desktop: bleak, Android: Kotlin BleBridge via Chaquopy)
+                           │
+                           └─ BLEProtocol (family-specific wire bytes: sp110e.py, triones.py, zengge.py, govee.py)
+```
+
+- One `BLEProtocol` dataclass per controller family. Each supplies GATT UUIDs, write type (with/without response), `encode_color` / `encode_power` functions, name prefixes for discovery, and an optional `init_writes` handshake sequence.
+- `BLEClient` is whole-strip only. `send_pixels()` averages incoming pixel arrays and emits one solid color per frame — none of these protocols support per-pixel streaming.
+- Discovery auto-detects the family via advertised name prefix first, falls back to service UUID matching. The detected family is returned on `DiscoveredDevice.ble_family` and preselected in the UI.
+- The settings modal lets users change the family after creation — wrong family → writes go to a characteristic the device ignores → strip stays dark.
+
+## Protocol Quirks
+
+### SP110E / SP108E (critical handshake)
+
+The controller **silently tears the GATT link down within ~1 second of connect** unless a two-write handshake arrives immediately:
+
+```
+Write  01 00                      → characteristic FFE2
+Write  01 B7 E3 D5                → characteristic FFE1
+```
+
+Without this, the first real write later hangs for 30 s because bleak thinks the link is up but the peripheral has already dropped it. We carry these writes in `PROTOCOL.init_writes` and execute them from `BLEClient.connect()` right after GATT open.
+
+Color frame is **4 bytes** (`RR GG BB 1E`), not 5 — the earlier implementation had a stray `0x00` padding byte that the device tolerated but isn't documented.
+
+Source: [mbullington's reverse-engineering gist](https://gist.github.com/mbullington/37957501a07ad065b67d4e8d39bfe012).
+
+### Triones / Zengge / Govee
+
+No init handshake required. Color frames and command bytes documented inline in each protocol module. Notable: Zengge and SP110E share service UUID `FFE0/FFE1`, so name-based identification is the only reliable way to tell them apart. In `_register_builtins()`, SP110E is registered first so it wins the `identify_family_by_service_uuids` tie by default — change this if the user base flips.
+
+## bleak + Windows WinRT Traps
+
+These bit us hard. All are now worked around, but future BLE work should keep them in mind.
+
+### 1. `asyncio.wait_for` hangs forever on WinRT
+
+`BleakClient.connect()` / `write_gatt_char()` wrap WinRT `IAsyncOperation`s. When asyncio tries to cancel them (as `wait_for` does on timeout), the WinRT task **never finishes cancelling**, so `wait_for` itself blocks forever while awaiting the cancellation. Symptom: log stops with no timeout error, process is alive but wedged.
+
+**Fix**: `_bounded_await()` in [ble_transport.py](../server/src/ledgrab/core/devices/ble_transport.py) uses `asyncio.wait()` instead, which returns on timeout without awaiting pending tasks. Orphans the hanging WinRT task but frees the caller.
+
+### 2. Connect by raw MAC string fails on Windows
+
+Passing `BleakClient("AA:BB:CC:DD:EE:FF")` makes WinRT guess the address type (public vs random static vs random resolvable). Guesses wrong → connect silently times out. Symptom: `TimeoutError: BLE connect to ... exceeded 10.0s` with no other signal.
+
+**Fix**: Always pre-scan with `BleakScanner.find_device_by_address()` and pass the returned `BLEDevice` object to `BleakClient`. Costs ~400 ms but makes connect reliable.
+
+### 3. Client-side fetch timeout too short for BLE target start
+
+The target-start endpoint does a ~5 s pre-scan + up to 10 s GATT connect + init handshake. Default `fetchWithAuth` has a 10 s timeout and 3× retry, so the UI was aborting and retrying concurrent `/start` requests into the server.
+
+**Fix**: `startTargetProcessing` overrides `timeout: 30000, retry: false`.
+
+### 4. `Start-Process -WindowStyle Hidden` from bash/WSL strips handles
+
+When `restart.ps1` is invoked from Git-Bash / WSL, `Start-Process` inherited handles cause the child uvicorn to exit immediately. Stream redirection fixes it.
+
+**Fix**: `restart.ps1` always uses `-RedirectStandardOutput`/`-RedirectStandardError` to a temp log. Failed startups dump the stderr tail to the caller so root cause is visible.
+
+## Vendor Lockdown (the dead end)
+
+Some controllers — notably the one we tested, advertising as `AlexTable` at `16:61:05:70:68:44` — **only accept connections from the vendor phone app**. Diagnostic sequence:
+
+| Test | Result | Meaning |
+| --- | --- | --- |
+| LedGrab `BleakClient.connect()` | 10 s timeout | Windows can't connect |
+| Windows "Bluetooth LE Explorer" | Hangs on connect | Same Windows stack as bleak — not our bug |
+| Phone **OS** Bluetooth Settings | Can't connect | Phone OS uses generic BLE stack — also fails |
+| Phone **LED Hue** app | Connects fine | Vendor app is the *only* working client |
+
+At this point, further Windows/bleak tweaks have no effect. The peripheral firmware rejects generic GATT connects and only stays connected when the LED Hue app emits its vendor-specific handshake. To unlock such a controller from LedGrab you'd need to:
+
+1. Enable **Developer Options → Bluetooth HCI snoop log** on Android.
+2. Reproduce the LED Hue flow (connect → color change → disconnect).
+3. `adb bugreport bugreport.zip`; extract `btsnoop_hci.log`.
+4. Open in Wireshark; identify the vendor handshake bytes written during connect.
+5. Add them to the protocol's `init_writes`.
+
+Alternatively, replace the BLE controller hardware with **WLED on ESP32** — $3, fully supported, vastly more capable.
+
+## Frontend
+
+- BLE family picker uses the project's shared `IconSelect` grid (project rule — see [CLAUDE.md](../CLAUDE.md): "NEVER use plain HTML `<select>`").
+- Registry in `device-discovery.ts` is keyed by element ID so both the add-device and settings modals get their own IconSelect instance. Helpers: `ensureBleFamilyIconSelect(selectId, onChange?)` / `destroyBleFamilyIconSelect(selectId)`.
+- Govee AES key row is conditionally visible: only shows when the selected family is `govee`.
+
+## HAOS Integration Pair
+
+The sister repo `ledgrab-haos-integration` had its own WebSocket auth bug that surfaced during this session — the integration still used the deprecated `?token=<key>` query param instead of the new first-message handshake. Fixed in [v0.2.1](https://git.dolgolyov-family.by/alexei.dolgolyov/ledgrab-haos-integration). Unrelated to BLE but shared debugging time.
+
+## Tests
+
+`server/tests/test_ble_protocols.py` and `server/tests/test_ble_client.py` use a `FakeTransport` that logs every write with its `char_uuid`, so protocol wire formats and the init handshake are unit-testable without hardware or bleak installed. New protocol additions should extend these.
+
+## Files
+
+- [ble_client.py](../server/src/ledgrab/core/devices/ble_client.py) — provider-facing class; runs init handshake on connect; reconnect backoff.
+- [ble_transport.py](../server/src/ledgrab/core/devices/ble_transport.py) — bleak desktop transport; `_bounded_await` helper; per-write char override.
+- [android_ble_transport.py](../server/src/ledgrab/core/devices/android_ble_transport.py) — Chaquopy/Kotlin transport; currently ignores `char_uuid` override (bridge binds a single write characteristic).
+- [ble_provider.py](../server/src/ledgrab/core/devices/ble_provider.py) — discovery, family detection, `set_color` / `set_power` short-lived sessions.
+- [ble_protocols/](../server/src/ledgrab/core/devices/ble_protocols/) — one file per family (pure byte-encoding functions, no BLE deps).
+- [BleBridge.kt](../android/app/src/main/java/com/ledgrab/android/BleBridge.kt) — Android-side BLE GATT wrapper exposed to Python via Chaquopy.

server/config/default_config.yaml

@@ -15,8 +15,8 @@ auth:
   # To enable LAN access, add one or more label: "api-key" entries below
   # and send `Authorization: Bearer <api-key>` with each request.
   # Generate secure keys: openssl rand -hex 32
-  api_keys: {}
+  api_keys:
-  #  dev: "replace-with-openssl-rand-hex-32"
+    dev: "development-key-change-in-production"
 
 # Storage paths default to ./data relative to the server's working directory.
 # Set LEDGRAB_DATA_DIR in the environment to point at a different data root

server/src/ledgrab/api/routes/output_targets_control.py

@@ -109,6 +109,7 @@ async def start_processing(
     manager: ProcessorManager = Depends(get_processor_manager),
 ):
     """Start processing for a output target."""
+    logger.info("Start processing requested for target %s", target_id)
     try:
         # Verify target exists in store
         target_store.get_target(target_id)

server/src/ledgrab/core/devices/android_ble_transport.py

@@ -120,14 +120,31 @@ class AndroidBLETransport:
         except Exception as exc:
             logger.warning("Android BLE disconnect of %s raised: %s", self._address, exc)
 
-    async def write(self, data: bytes) -> None:
+    async def write(self, data: bytes, char_uuid: Optional[str] = None) -> None:
-        """Write bytes to the configured characteristic.
+        """Write bytes to a characteristic on the connected peripheral.
 
         Serialised through an internal lock — BLE stacks do not tolerate
         overlapping writes on the same characteristic.
+
+        Args:
+            data: payload bytes.
+            char_uuid: currently unused on Android — the Kotlin BleBridge
+                binds a single "write characteristic" at connect time. If a
+                protocol's ``init_writes`` target a different characteristic
+                (e.g. SP110E's ``ffe2``), we log a warning and route the
+                write to the bound char anyway. This is lossy but keeps the
+                protocol layer compatible across backends until the bridge
+                grows per-write characteristic selection.
         """
         if not self.is_connected or self._handle is None:
             raise RuntimeError(f"Android BLE transport {self._address} not connected")
+        if char_uuid and char_uuid.lower() != self._write_char_uuid.lower():
+            logger.warning(
+                "Android BLE transport: init write to %s ignored (bridge is "
+                "bound to %s); vendor handshake may not complete",
+                char_uuid,
+                self._write_char_uuid,
+            )
         bridge = _bridge()
         handle = self._handle
         with_response = self._write_with_response

server/src/ledgrab/core/devices/ble_client.py

@@ -135,6 +135,31 @@ class BLEClient(LEDClient):
 
     async def connect(self) -> bool:
         await self._transport.connect()
+        # Some controllers (notably SP110E) require a vendor handshake within
+        # ~1 s of GATT connect or they silently tear the link down — which
+        # later surfaces as writes hanging until the Windows BLE stack
+        # times them out at 30 s. Run it here so send_pixels can assume the
+        # link is actually live.
+        if self._protocol.init_writes:
+            logger.debug(
+                "BLE init handshake: %d write(s) for %s",
+                len(self._protocol.init_writes),
+                self._protocol.family,
+            )
+            for char_uuid, payload in self._protocol.init_writes:
+                try:
+                    await self._transport.write(payload, char_uuid=char_uuid)
+                except Exception as exc:
+                    # Handshake failure is fatal — close and surface it so
+                    # the caller doesn't stream into a dead link.
+                    logger.warning(
+                        "BLE init write to %s on %s failed: %s",
+                        char_uuid,
+                        self._address,
+                        exc,
+                    )
+                    await self._transport.close()
+                    raise
         self._connected = True
         logger.info(
             "BLE client connected: address=%s family=%s", self._address, self._protocol.family

server/src/ledgrab/core/devices/ble_protocols/__init__.py

@@ -28,6 +28,12 @@ class BLEProtocol:
         encode_color: ``(r, g, b, brightness) -> bytes`` — frame setting a solid color.
         encode_power: ``on -> bytes`` — frame toggling power.
         name_prefixes: Advertisement-name prefixes that identify this family.
+        init_writes: Optional vendor-specific handshake performed right after
+            GATT connect. Each tuple is ``(char_uuid, payload)``. Some
+            controllers (notably SP110E) drop the link within ~1 s if this
+            handshake is missing, which is why our first real write hangs
+            for 30 s and then fails — the peripheral has already torn down
+            the connection, we just haven't noticed.
     """
 
     family: str
@@ -38,6 +44,7 @@ class BLEProtocol:
     encode_color: Callable[[int, int, int, int], bytes]
     encode_power: Callable[[bool], bytes]
     name_prefixes: Tuple[str, ...]
+    init_writes: Tuple[Tuple[str, bytes], ...] = ()
 
 
 _registry: Dict[str, BLEProtocol] = {}

server/src/ledgrab/core/devices/ble_protocols/sp110e.py

@@ -12,9 +12,9 @@ per-pixel frames. The BLE protocol exposes:
 
 So from LedGrab's perspective, SP110E is a whole-strip ambient controller.
 
-Frame format (5 bytes, big-endian):
+Frame format (4 bytes):
 
-    ``RR GG BB 00 CC``
+    ``RR GG BB CC``
 
 where ``CC`` is the command byte. Static-color command is ``0x1E`` (set
 "RGB" mode = whole-strip solid color from the RR GG BB payload). Power is
@@ -23,9 +23,19 @@ bytes ignored). Brightness is applied by the *caller* scaling the RGB
 triple — there is no separate brightness command for solid-color mode,
 which is simpler and lets LedGrab apply its own processing pipeline.
 
+Init handshake (critical):
+    The controller silently tears the GATT link down within ~1s if a
+    two-write handshake doesn't arrive immediately after connect. Omitting
+    it is indistinguishable from a successful connect on the Python side —
+    the first real write then hangs for 30s until Windows abandons the
+    queued packet. The handshake writes live in ``init_writes`` below and
+    are executed by ``BLEClient.connect()``.
+
 References:
     * https://github.com/Lehkeda/SP110E_controller (reverse-engineered)
     * https://github.com/sysofwan/ha-sp110e
+    * https://gist.github.com/mbullington/37957501a07ad065b67d4e8d39bfe012
+      (init handshake documentation)
 """
 
 from __future__ import annotations
@@ -34,6 +44,7 @@ from ledgrab.core.devices.ble_protocols import BLEProtocol
 
 _SERVICE_UUID = "0000ffe0-0000-1000-8000-00805f9b34fb"
 _WRITE_CHAR_UUID = "0000ffe1-0000-1000-8000-00805f9b34fb"
+_INIT_CHAR_UUID = "0000ffe2-0000-1000-8000-00805f9b34fb"
 
 _CMD_SET_COLOR = 0x1E
 _CMD_POWER_ON = 0xAA
@@ -62,13 +73,13 @@ def encode_color(r: int, g: int, b: int, brightness: int = 255) -> bytes:
         r = (r * brightness) // 255
         g = (g * brightness) // 255
         b = (b * brightness) // 255
-    return bytes((r, g, b, 0x00, _CMD_SET_COLOR))
+    return bytes((r, g, b, _CMD_SET_COLOR))
 
 
 def encode_power(on: bool) -> bytes:
     """Build a power on/off frame."""
     cmd = _CMD_POWER_ON if on else _CMD_POWER_OFF
-    return bytes((0x00, 0x00, 0x00, 0x00, cmd))
+    return bytes((0x00, 0x00, 0x00, cmd))
 
 
 PROTOCOL = BLEProtocol(
@@ -81,4 +92,10 @@ PROTOCOL = BLEProtocol(
     encode_color=encode_color,
     encode_power=encode_power,
     name_prefixes=("SP110E", "SP108E", "BLE-LED"),
+    # Vendor handshake — see module docstring. Without these writes the
+    # controller drops the GATT connection ~1s after it opens.
+    init_writes=(
+        (_INIT_CHAR_UUID, b"\x01\x00"),
+        (_WRITE_CHAR_UUID, b"\x01\xb7\xe3\xd5"),
+    ),
 )

server/src/ledgrab/core/devices/ble_transport.py

@@ -26,6 +26,31 @@ from ledgrab.utils.platform import is_android
 logger = get_logger(__name__)
 
 
+async def _bounded_await(coro, timeout: float, what: str):
+    """Like ``asyncio.wait_for`` but safe when the inner task can't be cancelled.
+
+    bleak's WinRT backend wraps WinRT ``IAsyncOperation`` s that ignore
+    asyncio cancellation. ``asyncio.wait_for`` responds to a timeout by
+    cancelling the inner task and then **awaiting its cancellation** — and
+    since the WinRT task never finishes cancelling, ``wait_for`` itself
+    hangs forever.
+
+    ``asyncio.wait`` with a timeout returns as soon as the timeout fires
+    regardless of whether pending tasks can actually be cancelled, so the
+    caller always regains control. The orphaned task is best-effort
+    cancelled and then abandoned.
+    """
+    task = asyncio.create_task(coro)
+    done, _pending = await asyncio.wait({task}, timeout=timeout)
+    if task not in done:
+        task.cancel()
+        raise asyncio.TimeoutError(f"{what} exceeded {timeout:.1f}s")
+    exc = task.exception()
+    if exc is not None:
+        raise exc
+    return task.result()
+
+
 def _bleak_available() -> bool:
     try:
         import bleak  # noqa: F401
@@ -127,15 +152,73 @@ class BLETransport:
         if self.is_connected:
             return
 
-        self._client = BleakClient(self._address, timeout=self._connect_timeout)
+        # Pre-scan for a fresh advertisement so bleak gets a BLEDevice (with
+        # address type + service UUIDs baked in) rather than just a raw MAC
+        # string. On Windows/WinRT, connecting by string fails/times out when
+        # the cached address type is wrong (common for cheap BLE LED chips
+        # that advertise as Random Static). Passing the BLEDevice avoids that.
+        from bleak import BleakScanner
+
+        scan_timeout = max(3.0, self._connect_timeout / 2)
+        logger.info("BLE pre-scan for %s (timeout=%.1fs)", self._address, scan_timeout)
+        try:
+            device = await _bounded_await(
+                BleakScanner.find_device_by_address(self._address, timeout=scan_timeout),
+                timeout=scan_timeout + 2.0,
+                what=f"BLE pre-scan for {self._address}",
+            )
+        except asyncio.TimeoutError as exc:
+            raise RuntimeError(str(exc)) from exc
+        except Exception as exc:
+            raise RuntimeError(
+                f"BLE pre-scan for {self._address} failed: {type(exc).__name__}: {exc}"
+            ) from exc
+        if device is None:
+            raise RuntimeError(
+                f"BLE device {self._address} not found in scan "
+                f"(timeout={scan_timeout}s) — is it powered on and in range?"
+            )
+        # Log everything bleak knows about this peripheral from the fresh
+        # advertisement. If connects keep timing out, look here for the
+        # address type (public vs random static) and advertised services —
+        # Windows WinRT is strict about matching those on connect.
+        details_summary = ""
+        try:
+            det = getattr(device, "details", None)
+            if det is not None:
+                details_summary = f" details={det!r}"
+        except Exception:
+            pass
+        logger.info(
+            "BLE pre-scan found %s name=%r rssi=%s%s — opening GATT",
+            self._address,
+            getattr(device, "name", None),
+            getattr(device, "rssi", None),
+            details_summary,
+        )
+
+        self._client = BleakClient(device, timeout=self._connect_timeout)
         try:
             # bleak's WinRT backend does not always respect the constructor
             # timeout — connect() can block 30s+ when the peripheral is gone.
-            # Wrap in wait_for so the Python-side bound is enforced.
+            # _bounded_await enforces the timeout on the Python side even when
-            await asyncio.wait_for(self._client.connect(), timeout=self._connect_timeout)
+            # the inner WinRT future refuses to be cancelled (wait_for hangs
+            # forever in that case — see _bounded_await docstring).
+            await _bounded_await(
+                self._client.connect(),
+                timeout=self._connect_timeout,
+                what=f"BLE connect to {self._address}",
+            )
         except Exception as exc:
             self._client = None
-            raise RuntimeError(f"Failed to connect to BLE device {self._address}: {exc}") from exc
+            # Many bleak/WinRT failures surface with an empty ``str(exc)`` —
+            # include the exception type so the log actually tells us which
+            # branch of "connect failed" we're in (timeout vs WinRT HRESULT
+            # vs BleakDeviceNotFoundError vs ...).
+            detail = f"{type(exc).__name__}: {exc}" if str(exc) else type(exc).__name__
+            raise RuntimeError(
+                f"Failed to connect to BLE device {self._address}: {detail}"
+            ) from exc
 
         logger.info("BLE connected to %s", self._address)
 
@@ -151,8 +234,8 @@ class BLETransport:
         except Exception as exc:
             logger.warning("BLE disconnect of %s raised: %s", self._address, exc)
 
-    async def write(self, data: bytes) -> None:
+    async def write(self, data: bytes, char_uuid: Optional[str] = None) -> None:
-        """Send bytes to the configured write characteristic.
+        """Send bytes to a GATT write characteristic.
 
         Serialised through an internal lock — BLE stacks do not like
         overlapping writes on the same GATT characteristic.
@@ -161,18 +244,25 @@ class BLETransport:
         its default 30s on the second write to certain cheap BLE LED chips.
         Timing out keeps the target's processing loop responsive.
 
+        Args:
+            data: payload bytes.
+            char_uuid: target characteristic. Defaults to the transport's
+                configured ``write_char_uuid`` (used for color/power frames);
+                pass an alternate UUID for vendor-specific init writes that
+                target a different characteristic (e.g. SP110E's ``ffe2``).
+
         Raises:
             RuntimeError: If not connected.
             TimeoutError: If the write does not complete within 2 seconds.
         """
         if not self.is_connected or self._client is None:
             raise RuntimeError(f"BLE transport {self._address} not connected")
+        target = char_uuid or self._write_char_uuid
         async with self._lock:
-            await asyncio.wait_for(
+            await _bounded_await(
-                self._client.write_gatt_char(
+                self._client.write_gatt_char(target, data, response=self._write_with_response),
-                    self._write_char_uuid, data, response=self._write_with_response
-                ),
                 timeout=2.0,
+                what=f"BLE write to {self._address}",
             )
 
 

server/src/ledgrab/static/js/features/device-discovery.ts

@@ -696,7 +696,8 @@ export function showAddDevice(presetType: any = null, cloneData: any = null) {
             const bleFamilyEl = document.getElementById('device-ble-family') as HTMLSelectElement;
             if (bleFamilyEl && cloneData.ble_family) {
                 bleFamilyEl.value = cloneData.ble_family;
-                if (_bleFamilyIconSelect) _bleFamilyIconSelect.setValue(cloneData.ble_family);
+                const iconSelect = _bleFamilyIconSelects['device-ble-family'];
+                if (iconSelect) iconSelect.setValue(cloneData.ble_family);
             }
             const goveeKeyEl = document.getElementById('device-ble-govee-key') as HTMLInputElement;
             if (goveeKeyEl && cloneData.ble_govee_key) goveeKeyEl.value = cloneData.ble_govee_key;
@@ -822,7 +823,8 @@ export function selectDiscoveredDevice(device: any) {
     if (isBleDevice(device.device_type) && device.ble_family) {
         const familyEl = document.getElementById('device-ble-family') as HTMLSelectElement;
         if (familyEl) familyEl.value = device.ble_family;
-        if (_bleFamilyIconSelect) _bleFamilyIconSelect.setValue(device.ble_family);
+        const iconSelect = _bleFamilyIconSelects['device-ble-family'];
+        if (iconSelect) iconSelect.setValue(device.ble_family);
         _updateBleGoveeKeyVisibility();
     }
     showToast(t('device.scan.selected'), 'info');
@@ -1286,26 +1288,43 @@ function _buildBleFamilyItems() {
     ];
 }
 
-let _bleFamilyIconSelect: any = null;
+// Parameterized by select element ID so both the add-device modal
+// (``device-ble-family``) and the settings modal (``settings-ble-family``)
+// get their own IconSelect instance — same pattern as DMX/SPI/etc.
+const _bleFamilyIconSelects: Record<string, any> = {};
 
-function _destroyBleFamilyIconSelect() {
+export function destroyBleFamilyIconSelect(selectId: string) {
-    if (_bleFamilyIconSelect) {
+    if (_bleFamilyIconSelects[selectId]) {
-        _bleFamilyIconSelect.destroy();
+        _bleFamilyIconSelects[selectId].destroy();
-        _bleFamilyIconSelect = null;
+        delete _bleFamilyIconSelects[selectId];
     }
 }
 
-function _ensureBleFamilyIconSelect() {
+export function ensureBleFamilyIconSelect(selectId: string, onChange?: () => void): any {
-    const sel = document.getElementById('device-ble-family') as HTMLSelectElement;
+    const sel = document.getElementById(selectId) as HTMLSelectElement | null;
-    if (!sel) return;
+    if (!sel) return null;
-    if (_bleFamilyIconSelect) {
+    if (_bleFamilyIconSelects[selectId]) {
-        _bleFamilyIconSelect.updateItems(_buildBleFamilyItems());
+        _bleFamilyIconSelects[selectId].updateItems(_buildBleFamilyItems());
-    } else {
+        return _bleFamilyIconSelects[selectId];
-        _bleFamilyIconSelect = new IconSelect({ target: sel, items: _buildBleFamilyItems(), columns: 2 } as any);
-        // Register once — native <select> change fires when IconSelect picks a value,
-        // which is what toggles the Govee key field.
-        sel.addEventListener('change', _updateBleGoveeKeyVisibility);
     }
+    _bleFamilyIconSelects[selectId] = new IconSelect({
+        target: sel,
+        items: _buildBleFamilyItems(),
+        columns: 2,
+    } as any);
+    if (onChange) {
+        sel.addEventListener('change', onChange);
+    }
+    return _bleFamilyIconSelects[selectId];
+}
+
+// Thin wrappers used by the add-device modal.
+function _destroyBleFamilyIconSelect() {
+    destroyBleFamilyIconSelect('device-ble-family');
+}
+
+function _ensureBleFamilyIconSelect() {
+    ensureBleFamilyIconSelect('device-ble-family', _updateBleGoveeKeyVisibility);
     _updateBleGoveeKeyVisibility();
 }
 

server/src/ledgrab/static/js/features/devices.ts

@@ -6,9 +6,9 @@ import {
     _deviceBrightnessCache, updateDeviceBrightness,
     csptCache,
 } from '../core/state.ts';
-import { API_BASE, getHeaders, fetchWithAuth, escapeHtml, isSerialDevice, isMockDevice, isMqttDevice, isWsDevice, isOpenrgbDevice, isDmxDevice, isGroupDevice } from '../core/api.ts';
+import { API_BASE, getHeaders, fetchWithAuth, escapeHtml, isSerialDevice, isMockDevice, isMqttDevice, isWsDevice, isOpenrgbDevice, isDmxDevice, isBleDevice, isGroupDevice } from '../core/api.ts';
 import { devicesCache } from '../core/state.ts';
-import { _fetchOpenrgbZones, _getCheckedZones, _splitOpenrgbZone, _getZoneMode, ensureDmxProtocolIconSelect, destroyDmxProtocolIconSelect, ensureSpiLedTypeIconSelect, destroySpiLedTypeIconSelect, ensureGameSenseDeviceTypeIconSelect, destroyGameSenseDeviceTypeIconSelect, addGroupChildSettingsWithId as _addGroupChildSettingsWithId, ensureGroupModeIconSelect, destroyGroupModeIconSelect } from './device-discovery.ts';
+import { _fetchOpenrgbZones, _getCheckedZones, _splitOpenrgbZone, _getZoneMode, ensureDmxProtocolIconSelect, destroyDmxProtocolIconSelect, ensureSpiLedTypeIconSelect, destroySpiLedTypeIconSelect, ensureGameSenseDeviceTypeIconSelect, destroyGameSenseDeviceTypeIconSelect, addGroupChildSettingsWithId as _addGroupChildSettingsWithId, ensureGroupModeIconSelect, destroyGroupModeIconSelect, ensureBleFamilyIconSelect, destroyBleFamilyIconSelect } from './device-discovery.ts';
 import { t } from '../core/i18n.ts';
 import { showToast, showConfirm, desktopFocus } from '../core/ui.ts';
 import { Modal } from '../core/modal.ts';
@@ -66,6 +66,8 @@ class DeviceSettingsModal extends Modal {
             dmxProtocol: (document.getElementById('settings-dmx-protocol') as HTMLSelectElement | null)?.value || 'artnet',
             dmxStartUniverse: (document.getElementById('settings-dmx-start-universe') as HTMLInputElement | null)?.value || '0',
             dmxStartChannel: (document.getElementById('settings-dmx-start-channel') as HTMLInputElement | null)?.value || '1',
+            bleFamily: (document.getElementById('settings-ble-family') as HTMLSelectElement | null)?.value || '',
+            bleGoveeKey: (document.getElementById('settings-ble-govee-key') as HTMLInputElement | null)?.value || '',
             csptId: (document.getElementById('settings-css-processing-template') as HTMLSelectElement | null)?.value || '',
         };
     }
@@ -443,6 +445,37 @@ export async function showSettings(deviceId: any) {
             if (dmxStartChannelGroup) (dmxStartChannelGroup as HTMLElement).style.display = 'none';
         }
 
+        // BLE-specific fields — exposed in the settings modal so the user
+        // can fix a wrong protocol family pick without deleting+recreating
+        // the device. Uses the shared IconSelect grid (project rule bans
+        // plain <select> pickers).
+        const bleFamilyGroup = document.getElementById('settings-ble-family-group');
+        const bleGoveeKeyGroup = document.getElementById('settings-ble-govee-key-group');
+        const bleFamilySelect = document.getElementById('settings-ble-family') as HTMLSelectElement | null;
+        const bleGoveeKeyInput = document.getElementById('settings-ble-govee-key') as HTMLInputElement | null;
+        const _toggleSettingsGoveeKeyVisibility = () => {
+            const family = bleFamilySelect?.value || '';
+            if (bleGoveeKeyGroup) {
+                (bleGoveeKeyGroup as HTMLElement).style.display = family === 'govee' ? '' : 'none';
+            }
+        };
+        if (isBleDevice(device.device_type)) {
+            if (bleFamilyGroup) (bleFamilyGroup as HTMLElement).style.display = '';
+            const currentFamily = device.ble_family || 'sp110e';
+            if (bleFamilySelect) bleFamilySelect.value = currentFamily;
+            if (bleGoveeKeyInput) bleGoveeKeyInput.value = device.ble_govee_key || '';
+            const iconSelect = ensureBleFamilyIconSelect(
+                'settings-ble-family',
+                _toggleSettingsGoveeKeyVisibility,
+            );
+            if (iconSelect) iconSelect.setValue(currentFamily);
+            _toggleSettingsGoveeKeyVisibility();
+        } else {
+            destroyBleFamilyIconSelect('settings-ble-family');
+            if (bleFamilyGroup) (bleFamilyGroup as HTMLElement).style.display = 'none';
+            if (bleGoveeKeyGroup) (bleGoveeKeyGroup as HTMLElement).style.display = 'none';
+        }
+
         // Group device fields
         const groupChildrenGroup = document.getElementById('settings-group-children-group');
         const groupModeGroup = document.getElementById('settings-group-mode-group');
@@ -499,7 +532,7 @@ export async function showSettings(deviceId: any) {
 }
 
 export function isSettingsDirty() { return settingsModal.isDirty(); }
-export function forceCloseDeviceSettingsModal() { if (_deviceTagsInput) { _deviceTagsInput.destroy(); _deviceTagsInput = null; } if (_settingsCsptEntitySelect) { _settingsCsptEntitySelect.destroy(); _settingsCsptEntitySelect = null; } settingsModal.forceClose(); }
+export function forceCloseDeviceSettingsModal() { if (_deviceTagsInput) { _deviceTagsInput.destroy(); _deviceTagsInput = null; } if (_settingsCsptEntitySelect) { _settingsCsptEntitySelect.destroy(); _settingsCsptEntitySelect = null; } destroyBleFamilyIconSelect('settings-ble-family'); settingsModal.forceClose(); }
 export function closeDeviceSettingsModal() { settingsModal.close(); }
 
 export async function saveDeviceSettings() {
@@ -542,6 +575,11 @@ export async function saveDeviceSettings() {
             body.dmx_start_universe = parseInt((document.getElementById('settings-dmx-start-universe') as HTMLInputElement | null)?.value || '0', 10);
             body.dmx_start_channel = parseInt((document.getElementById('settings-dmx-start-channel') as HTMLInputElement | null)?.value || '1', 10);
         }
+        if (isBleDevice(settingsModal.deviceType)) {
+            body.ble_family = (document.getElementById('settings-ble-family') as HTMLSelectElement | null)?.value || 'sp110e';
+            const goveeKey = (document.getElementById('settings-ble-govee-key') as HTMLInputElement | null)?.value?.trim() || '';
+            body.ble_govee_key = goveeKey;
+        }
         if (isGroup) {
             const childRows = document.querySelectorAll('#settings-group-children-list .group-child-row') as NodeListOf<HTMLElement>;
             body.group_device_ids = Array.from(childRows).map(r => r.dataset.deviceId || '').filter(v => v !== '');

server/src/ledgrab/static/js/features/targets.ts

@@ -1091,8 +1091,15 @@ async function _targetAction(action: any) {
 
 export async function startTargetProcessing(targetId: any) {
     await _targetAction(async () => {
+        // Start can take a while for devices whose connect path includes a
+        // BLE pre-scan + GATT connect (seconds, not milliseconds). Don't
+        // retry — a duplicate /start while the first one is in-flight just
+        // piles work on the server. 30s matches the server's own BLE
+        // connect budget with headroom.
         const response = await fetchWithAuth(`/output-targets/${targetId}/start`, {
             method: 'POST',
+            timeout: 30000,
+            retry: false,
         });
         if (response.ok) {
             showToast(t('device.started'), 'success');

server/src/ledgrab/templates/modals/device-settings.html

@@ -132,6 +132,31 @@
                     <input type="number" id="settings-dmx-start-channel" min="1" max="512" value="1">
                 </div>
 
+                <!-- BLE LED Controller fields -->
+                <div class="form-group" id="settings-ble-family-group" style="display: none;">
+                    <div class="label-row">
+                        <label for="settings-ble-family" data-i18n="device.ble.family">Protocol Family:</label>
+                        <button type="button" class="hint-toggle" onclick="toggleHint(this)" title="?">?</button>
+                    </div>
+                    <small class="input-hint" style="display:none" data-i18n="device.ble.family.hint">Which BLE protocol your controller speaks. Match the phone app you normally use.</small>
+                    <select id="settings-ble-family">
+                        <option value="sp110e">SP110E / SP108E</option>
+                        <option value="triones">Triones / HappyLighting / LEDnet</option>
+                        <option value="zengge">Zengge / iLightsIn</option>
+                        <option value="govee">Govee (experimental)</option>
+                    </select>
+                </div>
+                <div class="form-group" id="settings-ble-govee-key-group" style="display: none;">
+                    <div class="label-row">
+                        <label for="settings-ble-govee-key" data-i18n="device.ble.govee_key">Govee AES Key (hex):</label>
+                        <button type="button" class="hint-toggle" onclick="toggleHint(this)" title="?">?</button>
+                    </div>
+                    <small class="input-hint" style="display:none" data-i18n="device.ble.govee_key.hint">Optional. Newer Govee firmware needs a per-model AES key — leave blank for older firmware.</small>
+                    <input type="text" id="settings-ble-govee-key"
+                           data-i18n-placeholder="device.ble.govee_key.placeholder"
+                           placeholder="32 hex digits, e.g. 0102…1f20">
+                </div>
+
                 <!-- Group device fields -->
                 <div class="form-group" id="settings-group-children-group" style="display: none;">
                     <div class="label-row">

server/tests/test_ble_client.py

@@ -19,6 +19,9 @@ class FakeTransport:
 
     def __init__(self, *_, **__):
         self.writes: List[bytes] = []
+        # Full log preserving char_uuid alongside payload so tests can
+        # assert on the SP110E vendor handshake's characteristic targets.
+        self.writes_detailed: List[tuple] = []
         self._connected = False
 
     @property
@@ -31,10 +34,17 @@ class FakeTransport:
     async def close(self) -> None:
         self._connected = False
 
-    async def write(self, data: bytes) -> None:
+    async def write(self, data: bytes, char_uuid: str | None = None) -> None:
         if not self._connected:
             raise RuntimeError("fake transport not connected")
         self.writes.append(data)
+        self.writes_detailed.append((char_uuid, data))
+
+    @property
+    def color_writes(self) -> list:
+        """Writes that went to the protocol's default write characteristic
+        (i.e. not the init handshake). Most existing tests want this view."""
+        return [data for (char_uuid, data) in self.writes_detailed if char_uuid is None]
 
 
 @pytest.fixture
@@ -79,14 +89,27 @@ class TestBLEClientLifecycle:
         await client.connect()
         assert client.is_connected
 
+    @pytest.mark.asyncio
+    async def test_connect_runs_vendor_init_handshake(self, fake_transport_cls):
+        # SP110E requires a two-write handshake on connect or the GATT
+        # link silently drops — verify both writes actually go out.
+        client = BLEClient("ble://AA:BB:CC", ble_family="sp110e")
+        await client.connect()
+        init_writes = client._transport.writes_detailed
+        assert len(init_writes) == 2
+        (char_a, payload_a), (char_b, payload_b) = init_writes
+        assert "ffe2" in char_a and payload_a == b"\x01\x00"
+        assert "ffe1" in char_b and payload_b == b"\x01\xb7\xe3\xd5"
+
     @pytest.mark.asyncio
     async def test_close_does_not_send_power_off(self, fake_transport_cls):
         client = BLEClient("ble://AA:BB:CC", ble_family="sp110e")
         await client.connect()
         await client.close()
         # Strip is left in whatever state it's in — rapid power toggles on
-        # connect/close cause BLE stack hangs on Windows.
+        # connect/close cause BLE stack hangs on Windows. (Only check color
-        assert bytes((0, 0, 0, 0, 0xAB)) not in client._transport.writes
+        # writes since the vendor init handshake is unrelated.)
+        assert bytes((0, 0, 0, 0xAB)) not in client._transport.color_writes
 
     @pytest.mark.asyncio
     async def test_unknown_family_raises(self):
@@ -101,10 +124,12 @@ class TestBLEClientSendPixels:
         await client.connect()
         ok = await client.send_pixels([(255, 0, 0), (0, 0, 0)], brightness=255)
         assert ok
-        assert len(client._transport.writes) == 1
+        # Color frames go to the default write characteristic; init-handshake
-        frame = client._transport.writes[0]
+        # writes have their own char_uuid and don't count here.
-        # Averaged to (127, 0, 0), SP110E trailer 00 1E
+        color_writes = client._transport.color_writes
-        assert frame == bytes((127, 0, 0, 0, 0x1E))
+        assert len(color_writes) == 1
+        # Averaged to (127, 0, 0), SP110E 4-byte frame with 0x1E cmd tail.
+        assert color_writes[0] == bytes((127, 0, 0, 0x1E))
 
     @pytest.mark.asyncio
     async def test_duplicate_frames_are_dropped(self, fake_transport_cls):
@@ -113,7 +138,7 @@ class TestBLEClientSendPixels:
         await client.send_pixels([(10, 20, 30)])
         await client.send_pixels([(10, 20, 30)])
         await client.send_pixels([(10, 20, 30)])
-        assert len(client._transport.writes) == 1
+        assert len(client._transport.color_writes) == 1
 
     @pytest.mark.asyncio
     async def test_send_returns_false_when_disconnected(self, fake_transport_cls):
@@ -127,6 +152,7 @@ class TestBLEClientSendPixels:
         client = BLEClient("ble://AA:BB:CC", ble_family="triones")
         await client.connect()
         await client.send_pixels([(100, 150, 200)])
+        # Triones has no init handshake — first write is the color frame.
         frame = client._transport.writes[0]
         assert frame == bytes((0x7E, 0x07, 0x05, 0x03, 100, 150, 200, 0x10, 0xEF))
 
@@ -209,5 +235,7 @@ class TestGoveeAESEncryption:
         assert client._aes_key is None
         await client.connect()
         await client.send_pixels([(100, 100, 100)])
-        # SP110E frame is 5 bytes, not encrypted.
+        # SP110E color frame is 4 bytes (RR GG BB CMD), not encrypted.
-        assert len(client._transport.writes[0]) == 5
+        color_writes = client._transport.color_writes
+        assert len(color_writes) == 1
+        assert len(color_writes[0]) == 4

server/tests/test_ble_protocols.py

@@ -18,24 +18,34 @@ from ledgrab.core.devices.ble_protocols import (
 
 
 class TestSP110E:
-    def test_color_frame_is_five_bytes_with_cmd_tail(self):
+    def test_color_frame_is_four_bytes_with_cmd_tail(self):
         frame = sp110e.encode_color(255, 128, 64)
-        assert frame == bytes((255, 128, 64, 0x00, 0x1E))
+        assert frame == bytes((255, 128, 64, 0x1E))
 
     def test_brightness_scales_rgb(self):
         frame = sp110e.encode_color(200, 200, 200, brightness=128)
         # 200 * 128 // 255 == 100
-        assert frame == bytes((100, 100, 100, 0x00, 0x1E))
+        assert frame == bytes((100, 100, 100, 0x1E))
 
     def test_brightness_255_is_passthrough(self):
-        assert sp110e.encode_color(1, 2, 3, 255) == bytes((1, 2, 3, 0x00, 0x1E))
+        assert sp110e.encode_color(1, 2, 3, 255) == bytes((1, 2, 3, 0x1E))
 
     def test_clamps_out_of_range(self):
-        assert sp110e.encode_color(-5, 300, 128) == bytes((0, 255, 128, 0x00, 0x1E))
+        assert sp110e.encode_color(-5, 300, 128) == bytes((0, 255, 128, 0x1E))
 
     def test_power_frames(self):
-        assert sp110e.encode_power(True) == bytes((0, 0, 0, 0, 0xAA))
+        assert sp110e.encode_power(True) == bytes((0, 0, 0, 0xAA))
-        assert sp110e.encode_power(False) == bytes((0, 0, 0, 0, 0xAB))
+        assert sp110e.encode_power(False) == bytes((0, 0, 0, 0xAB))
+
+    def test_init_handshake_is_defined(self):
+        # SP110E silently drops the GATT link within ~1s of connect unless
+        # this two-write handshake arrives — see module docstring.
+        assert len(sp110e.PROTOCOL.init_writes) == 2
+        (ffe2_uuid, ffe2_payload), (ffe1_uuid, ffe1_payload) = sp110e.PROTOCOL.init_writes
+        assert ffe2_uuid.endswith("ffe2-0000-1000-8000-00805f9b34fb") or "ffe2" in ffe2_uuid
+        assert ffe2_payload == b"\x01\x00"
+        assert ffe1_uuid == sp110e.PROTOCOL.write_char_uuid
+        assert ffe1_payload == b"\x01\xb7\xe3\xd5"
 
 
 class TestTriones: