chore: reset NIGHTLY_CHANGELOG.md after nightly-0.9.30-nightly10 release

Discovered while testing the smart-plug power-switch feature: unplugging
the printer left the dashboard stuck showing it as online/"ready"
indefinitely. Live-tested against a real printer to isolate two
independent, compounding causes:

1. The MQTT socket had no TCP keepalive. A connection killed without a
   clean TCP close (unplugged, not a graceful shutdown) looks alive to the
   OS for as long as its default dead-connection timeout - 15+ minutes on
   Linux - since a send on a half-open connection is buffered by the
   kernel and doesn't fail immediately. Fixed with SO_KEEPALIVE (short
   idle/interval/count) plus TCP_USER_TIMEOUT, since keepalive probes alone
   only fire on an idle connection - verified live that a printer
   disappearing while a send was still in flight (the common case, since
   the poll loop sends every few seconds) instead falls back to the far
   slower normal TCP retransmission timer, which keepalive settings don't
   affect at all.

2. Even after the socket was correctly detected as dead, the status poll
   loop could hang indefinitely inside publish() waiting for a reconnect
   attempt already running on the MQTT reader thread (the Issue #105
   reconnect-lock serialization), and therefore never reached the
   is_connected() check that flips kobra_state to "offline". _reconnect()
   now takes wait_if_in_progress/persist flags so the poll loop's call
   returns immediately with at most one attempt instead of blocking
   through someone else's multi-minute backoff loop - persistent retrying
   stays the reader thread's job.

A disconnected printer is now detected and reflected on the dashboard
within about 15 seconds. Live-verified across repeated disconnect/
reconnect cycles that no sockets, threads, or file descriptors are left
behind (checked via /proc/<pid>/fd and /proc/<pid>/task) - the transient
FIN-WAIT-2 entries seen while the printer's TLS service is still booting
belong to the kernel's own connection teardown, not to processes held by
the bridge, and clear on their own.

API.md

@@ -0,0 +1,256 @@
+# KX-Bridge HTTP API Reference
+
+This document lists the HTTP and WebSocket surface exposed by
+`kobrax_moonraker_bridge.py`. It is a reference for integrators and
+plugin authors, not a tutorial — for day-to-day usage of the bridge see
+[MANUAL.md](MANUAL.md), and for setup see [README.md](README.md).
+
+The API has two distinct parts:
+
+1. **Moonraker-compatible surface** — a subset of the real
+   [Moonraker](https://moonraker.readthedocs.io/) HTTP + WebSocket API,
+   implemented just far enough to make Mainsail, Fluidd, OrcaSlicer, and
+   `moonraker-obico` work against the Kobra X. **This is not a full
+   Moonraker implementation** — many real Moonraker endpoints/methods do
+   not exist here, and some responses are static stubs that exist only
+   to stop a client from erroring/looping (noted below).
+2. **Bridge-specific surface** — `/api/...` and `/kx/...` endpoints for
+   things Moonraker has no concept of: multi-printer management, AMS/ACE
+   filament control, the GCode store, custom filament profile import,
+   Spoolman, and the smart-plug power switch.
+
+## Security
+
+**There is no authentication on any endpoint.** The bridge is designed
+to be run on a trusted local network only. Do not expose port `7125`
+(or any additional per-printer port) to the internet — anyone who can
+reach the port can control the printer, read/delete files, and read
+`/kx/printers` credentials indirectly through bridge behavior. `/access/api_key`
+returns a hardcoded dummy value purely so `moonraker-obico` doesn't warn;
+it is not a real credential.
+
+CORS is enabled (`_json_cors` / `handle_kx_options` add
+`Access-Control-Allow-*` headers and answer `OPTIONS` with 204) so the
+Web UI can call sibling bridge instances directly in multi-printer setups.
+
+---
+
+## Moonraker-compatible endpoints (HTTP)
+
+All responses follow Moonraker's `{"result": {...}}` envelope unless noted.
+
+| Method | Path | Purpose | Notes |
+|---|---|---|---|
+| GET | `/server/info` | Server/klippy status | Always reports `klippy_connected: true`, `klippy_state: "ready"` |
+| GET | `/printer/info` | Printer identity | Static hostname/paths; `software_version` from `KLIPPER_VERSION` |
+| GET | `/machine/system_info` | System info stub | Mostly static/placeholder fields |
+| GET | `/printer/objects/list` | List available printer objects | Keys of `_build_printer_objects()` |
+| GET | `/printer/objects/query?objects=...` | Query object status | Comma-separated `objects` query param, or bare query keys |
+| GET`/POST` | `/printer/objects/subscribe` | Subscribe (HTTP polling variant) | Returns full status snapshot immediately |
+| GET | `/server/files/list` | List gcode files | Only returns the single currently-tracked file (if any) |
+| GET | `/server/files/metadata?filename=...` | File metadata (layers, est. time, etc.) | Shared logic with WS `server.files.metadata`; falls back to GCode store / buried-report cache |
+| POST | `/server/files/upload` | Upload a gcode file (multipart) | Same handler as `/api/files/local` |
+| POST | `/printer/print/start?filename=...` | Start a print | Body may include `filament_assignments`, `excluded_objects`, `auto_leveling` |
+| POST | `/printer/print/pause` | Pause current print | |
+| POST | `/printer/print/resume` | Resume current print | |
+| POST | `/printer/print/cancel` | Cancel current print | |
+| GET | `/access/api_key` | Dummy API key | No real auth exists |
+| GET | `/machine/update/status` | Update-manager stub | Always `busy: false`, empty `version_info` |
+| GET | `/server/history/list?limit=` | Print job history | Backed by the bridge's own GCodeStore/job DB |
+| GET | `/server/webcams/list` | Webcam descriptor | Rewrites `localhost`/`127.0.0.1` Host header to the bridge's LAN IP so remote Obico/Mainsail instances get a reachable URL |
+| POST | `/printer/gcode/script` | Execute a (very limited) gcode command | See `_exec_gcode_script`; not a general gcode interpreter |
+| GET | `/server/database/item?namespace=&key=` | Moonraker "database" KV read | Real payload only for `lane_data` (AMS/filament sync for OrcaSlicer); stub/empty responses for `AFC`, `afc-install`, `happy_hare`, `mainsail`; in-memory KV for `obico` |
+| POST | `/server/database/item` | Moonraker "database" KV write | In-memory only (not persisted across restarts); used by `moonraker-obico` for its own settings |
+| GET | `/server/database/list` | List KV namespaces | Static: `["lane_data", "mainsail", "obico"]` |
+
+### OctoPrint-compatibility shim
+
+| Method | Path | Purpose |
+|---|---|---|
+| GET | `/api/version` | OctoPrint-style version probe (some tools check this instead of Moonraker) |
+| POST | `/api/files/local`, `/api/files/{path}` | Alias for the same multipart upload handler as `/server/files/upload` |
+
+### WebSocket JSON-RPC (`/websocket`)
+
+Moonraker's JSON-RPC 2.0 protocol over a single `/websocket` endpoint. On
+connect the bridge immediately pushes `notify_klippy_ready` and then
+periodic `notify_status_update` notifications. Supported `method` values:
+
+| Method | Purpose |
+|---|---|
+| `printer.info` / `printer_info` | Same payload as `/printer/info` |
+| `server.info` / `server_info` | Same payload as `/server/info` |
+| `printer.objects.list` | Same as HTTP equivalent |
+| `printer.objects.query` / `printer.objects.get` | Object status by requested keys |
+| `printer.objects.subscribe` | Returns a status snapshot (no real push subscription semantics — status pushes happen automatically via `notify_status_update`) |
+| `printer.print.start` | `params.filename` |
+| `printer.print.pause` / `.resume` / `.cancel` | |
+| `machine.system_info` | Minimal stub |
+| `server.files.list` | Always returns `[]` over WS (unlike the HTTP version) |
+| `printer.gcode.script` | `params.script`, same limited executor as the HTTP endpoint |
+| `server.connection.identify` | Returns a dummy `connection_id` for Obico's handshake |
+| `connection.register_remote_method` | Accepted and ignored (Obico registers a remote-event callback) |
+| `server.webcams.list` | Same shape as HTTP, using the bridge's own LAN IP |
+| `server.history.list` | Job history, same source as `/server/history/list` |
+| `machine.update.status` | Stub |
+| `server.files.metadata` | Same logic as `/server/files/metadata` |
+
+Any other method is logged and answered with an empty `result: {}` — it
+does not error, to avoid breaking clients that probe for optional
+methods.
+
+---
+
+## Bridge-specific endpoints
+
+All `/kx/...` (and most `/api/...`) responses use `{"result": ...}` on
+success and `{"error": "..."}` with a 4xx/5xx status on failure, except
+where noted.
+
+### Printer control (`/api/...`)
+
+| Method | Path | Purpose | Body / Query |
+|---|---|---|---|
+| POST | `/api/light` | Toggle chamber light | `{on, brightness}` |
+| POST | `/api/fan` | Set part-cooling fan speed | `{speed}` (0–100) |
+| POST | `/api/connect` | Manually (re)connect the MQTT client | — |
+| POST | `/api/disconnect` | Manually disconnect | — |
+| POST | `/api/restart` | Restart the bridge process | — |
+| POST | `/api/speed` | Set print speed mode | `{mode}` (int) |
+| POST | `/api/axis` | Jog an axis, or `{"action":"turnoff"}` to disable steppers | `{axis, move_type, distance}` |
+| POST | `/api/temperature` | Set nozzle/bed target temps | `{nozzle?, bed?}`; uses a different MQTT path mid-print vs. idle |
+| GET | `/api/state` | Full dashboard status snapshot | Primary polling endpoint used by the Web UI |
+| GET | `/api/camera` | Current camera stream URL | |
+| GET | `/api/camera/stream` | MJPEG live view | `multipart/x-mixed-replace`, fed from a shared ffmpeg fanout |
+| GET | `/api/camera/h264` | Raw H.264 stream (for Obico) | |
+| GET | `/api/camera/snapshot` | Last cached JPEG frame | Instant, served from RAM |
+| POST | `/api/camera/start` / `/api/camera/stop` / `/api/camera/reset` | Camera lifecycle control | `reset` clears the 429 backoff and restarts ffmpeg |
+| GET | `/api/settings` | Read current config.ini-backed settings | |
+| POST | `/api/settings` | Write settings, then restart the bridge | See config fields below |
+| GET | `/api/update/check` | Check Gitea releases for a newer version | Branches on nightly/dev/stable channel |
+| POST | `/api/update/apply` | Self-update (non-Docker builds only) | `{tag}` |
+| POST | `/api/file_ready/clear` | Dismiss the "file ready to print" banner/dialog state | |
+| GET | `/api/log/stream` | Server-Sent Events log tail | |
+| GET | `/api/log/download` | Download buffered log as plaintext | |
+| GET | `/serve/{filename}` | Internal file server used to hand the printer a URL to fetch gcode from | Not meant for direct browser use |
+
+**`/api/settings` fields** (POST body, all optional — merges into
+existing config.ini): `printer_ip`, `mqtt_port`, `username`, `password`,
+`mode_id`, `device_id`, `power_on_url`, `power_off_url`,
+`power_status_url`, `default_ams_slot`, `auto_leveling`,
+`vibration_compensation`, `camera_on_print`, `web_upload_warning`,
+`print_start_dialog`, `poll_interval`, `verbose_http_log`,
+`printer_name`, `spoolman_server`, `spoolman_sync_rate`,
+`ace_dry_presets`.
+
+### AMS / ACE filament control (`/api/...`)
+
+| Method | Path | Purpose | Body |
+|---|---|---|---|
+| POST | `/api/ams/set_slot` | Set material type + color for a slot | `{index, type, color:[r,g,b]}` |
+| POST | `/api/ams/feed` | Feed filament in/out | `{slot_index, type}` (1=feed in, 2=feed out) |
+| POST | `/api/ace/auto_feed` | Toggle auto-feed for an ACE unit | `{ace_id, on}` |
+| POST | `/api/ace/dry` | Start/stop the ACE dryer | `{action: "start"|"stop", ace_id?, target_temp?, duration?}` |
+
+### GCode store (bridge-managed uploads) (`/kx/files...`)
+
+| Method | Path | Purpose |
+|---|---|---|
+| GET | `/kx/files` | List files the bridge has stored, with last-print status/duration |
+| DELETE | `/kx/files/{file_id}` | Delete a stored file |
+| GET | `/kx/files/{file_id}/download` | Download a stored file |
+| POST | `/kx/files/{file_id}/verify` | Clear the "web upload, unverified" flag |
+| GET | `/kx/files/{id}/objects` | Print-object list + SVG preview (for the pre-print skip feature) |
+| GET | `/kx/history?limit=&offset=` | Paginated print job history |
+
+### Files on the printer's own storage (`/kx/printer-files...`)
+
+Distinct from the GCode store above — these list/manage files that live
+on the printer's internal storage (e.g. printed directly from Anycubic
+Slicer Next, bypassing the bridge).
+
+| Method | Path | Purpose |
+|---|---|---|
+| GET | `/kx/printer-files` | List files via the printer's `file/listLocal` MQTT action |
+| POST | `/kx/printer-files/delete` | Delete one or more files: `{"filenames": [...]}` (single endpoint for single + bulk delete) |
+| GET | `/kx/printer-files/{filename}/thumbnail` | Fetch (and cache) a file's embedded gcode thumbnail via `file/fileDetails` |
+
+### Printing (`/kx/print`)
+
+| Method | Path | Purpose | Body |
+|---|---|---|---|
+| POST | `/kx/print` | Start a print from a stored GCode-store file | `{file_id, filament_assignments?, excluded_objects?, auto_leveling?}` |
+
+`filament_assignments` is `[{slot_index, material, color_hex}, ...]`; if
+omitted, all currently-occupied AMS slots are auto-mapped.
+
+### Pre-print / mid-print object skip (`/kx/skip...`)
+
+| Method | Path | Purpose |
+|---|---|---|
+| POST | `/kx/skip` | Skip named objects mid-print: `{"names": [...]}` |
+| POST | `/kx/skip/query` | Re-request the object list from the printer and return merged skip state |
+| GET | `/kx/skip/state` | Current skip state (object list, already-skipped names, SVG, filename) |
+
+### Filament profiles (`/kx/filament/...`)
+
+| Method | Path | Purpose | Body / Query |
+|---|---|---|---|
+| GET | `/kx/filament/slots` | Current AMS slot contents + any user profile override | |
+| GET | `/kx/filament/profiles?type=&vendor=` | Curated OrcaSlicer filament profile catalog (system + user-imported) | Optional filters |
+| GET | `/kx/filament/profiles/user` | User-imported profiles only (for the settings management list) | |
+| POST | `/kx/filament/profiles/user` | Import profiles from a ZIP or `.json` file(s) (multipart) | Multipart field `file`/`files`/`upload`; ZIP entries or bare `.json`, parsed via `orca_filaments.parse_profile_bytes` |
+| DELETE | `/kx/filament/profiles/user?vendor=&name=` | Delete one user profile (both params) or all (no params) | |
+| POST | `/kx/filament/slots/{idx}/profile` | Assign (or clear) a fixed profile override for one AMS slot | `{vendor, name}`; empty strings clear the mapping. Selector is `(vendor, name)`, not `id` — IDs are not unique across the Orca profile catalog |
+| GET`/POST` | `/kx/filament/visible_vendors` | Get/set the vendor visibility filter for the slot profile dropdown | POST body `{"vendors": [...]}`; empty list = show all |
+
+### Spoolman integration (`/kx/spoolman/...`)
+
+| Method | Path | Purpose |
+|---|---|---|
+| GET | `/kx/spoolman/status` | Whether Spoolman is configured/reachable, server URL, sync rate, current slot→spool map |
+| GET | `/kx/spoolman/spools` | Proxied list of spools from the configured Spoolman server |
+| POST | `/kx/spoolman/active-spool` | Assign spool IDs to AMS slots: `{"slot_map": {"0": 42, "2": 17}}` (AMS slot index → Spoolman spool ID) |
+
+### Multi-printer management (`/kx/printers...`)
+
+| Method | Path | Purpose | Body |
+|---|---|---|---|
+| GET | `/kx/printers` | List all configured printers with online-ish metadata | |
+| POST | `/kx/printers/add` | Add a printer by IP (credentials auto-fetched from the printer) | `{printer_ip, name?}` — triggers a bridge restart |
+| DELETE | `/kx/printers/{pid}` | Remove a printer from config; renumbers remaining `[printer_N]` sections | — triggers a bridge restart |
+| POST | `/kx/printers/{pid}/power` | Toggle an external smart plug (Tasmota-style) for a printer | `{"action": "on"|"off"}` |
+| GET | `/kx/printers/{pid}/power-status` | Query the smart plug's current on/off state | |
+
+**Power switch is not the printer's own power state** — it's a plain
+`GET` fired at a user-configured `power_on_url` / `power_off_url` /
+`power_status_url` (e.g. a Tasmota `cmnd=Power%20on` URL). See
+[MANUAL.md](MANUAL.md#power-switch-feature) for details.
+It only exists for printers where `power_on_url` or `power_off_url` is
+set in config; `/kx/printers` exposes this as `has_power_control`.
+
+### Misc
+
+| Method | Path | Purpose |
+|---|---|---|
+| GET | `/kx/ui/{name}` | Serves theme assets (JS/CSS/vendored libs) and translation JSON files under the active UI theme |
+| GET | `/` , `/printer{N}` | Serves the Web UI (index.html with CSS/JS inlined for embedded-webview compatibility, e.g. OrcaSlicer's device tab) |
+| GET | `/favicon.ico` | Favicon |
+
+---
+
+## Response conventions
+
+- Moonraker-compatible endpoints wrap results as `{"result": {...}}` (or
+  `{"error": {"code": ..., "message": ...}}` for the `/server/database/*`
+  404 case) to match the real Moonraker schema.
+- Bridge-specific `/api/...` and `/kx/...` endpoints generally return
+  `{"result": ...}` on success and `{"error": "message"}` with a
+  non-2xx HTTP status on failure — but this is not universal; check the
+  handler in `kobrax_moonraker_bridge.py` if exact shape matters (route
+  registrations are near the end of the file, search for
+  `r.add_get(`/`r.add_post(`/`r.add_delete(`).
+- Endpoints that trigger a config write (`/api/settings`,
+  `/kx/printers/add`, `/kx/printers/{pid}` DELETE) restart the whole
+  bridge process shortly after responding — clients should expect a
+  brief connection drop.

MANUAL.md

@@ -0,0 +1,330 @@
+# KX-Bridge User Manual
+
+This is a day-to-day how-to guide for using KX-Bridge once it's running.
+For installing/updating the bridge itself, see [README.md](README.md).
+For the HTTP/WebSocket API (developers, plugin authors, integrators), see
+[API.md](API.md).
+
+---
+
+## Getting Started
+
+### Install and start the bridge
+
+Follow the Quick Start in [README.md](README.md#-quick-start) — the short
+version is:
+
+```bash
+docker compose up -d
+```
+
+then open `http://BRIDGE-IP:7125` in a browser.
+
+### Connect to your printer for the first time
+
+1. On the printer's display: **Settings → Enable LAN mode**.
+2. In the bridge Web UI, the **Printers** tab shows **"+ Add printer"** on
+   first start. Click it, enter the printer's IP address, and confirm —
+   username, password and device ID are fetched from the printer and
+   decrypted automatically. No manual credential entry needed.
+3. Click **⚡ Connect** in the top-right corner to open the
+   connection. The status badge next to it shows the current printer state
+   (Standby, Printing, …).
+
+### Connect OrcaSlicer
+
+In OrcaSlicer, set the printer's connection type to **Moonraker** and enter
+`http://BRIDGE-IP:7125` as the host (full URL including `http://` and the
+port). See the [Recommended Slicer](README.md#-recommended-slicer) section
+of the README for the patched OrcaSlicer-KX build with proper per-slot
+filament matching.
+
+---
+
+## Dashboard Overview
+
+The Dashboard is the main screen and is made of movable/resizable tiles
+("cards"). Click **🖉 Customize dashboard** (top right
+of the dashboard) to enter edit mode: drag tiles to reorder, drag corners to
+resize, then save the arrangement as a named **preset** via the preset
+dropdown, or reset back to the default layout.
+
+The default tiles are:
+
+- **Camera** — live view from the printer's camera, with a light toggle and
+  a play/stop button. A small ↺ reset button appears if the stream needs to
+  reconnect (e.g. after a 429 rate-limit from the printer).
+- **Progress** — print percentage, a thumbnail of the current file, current
+  layer, current Z-height, elapsed/remaining time, and the file name.
+  While printing this card shows **Pause**, **Objects** (partial cancel, see
+  [Printing](#printing)) and **Stop** buttons. When a file is loaded but not
+  yet started, it instead shows **Print**, **Assign Slots**, and **Clear**.
+- **Temperatures** — current and target nozzle/bed temperature with
+  progress bars, quick "Set"/"Off" controls, and a rolling history chart of
+  the last 60 readings.
+- **Axis control** — jog buttons for X/Y/Z, adjustable step size (0.1 / 1 /
+  5 / 10 mm or a custom value), Home XY, Home Z, Home All, and Motors Off.
+- **Print Speed** — three presets (Quiet / Normal / Sport) matching the
+  printer's own speed modes.
+- **Fan** — a slider plus quick buttons (0/25/50/75/100%) for the part-cooling
+  fan.
+- **Filament / AMS** — one tile per AMS/ACE slot showing assigned material,
+  color and (if configured) the mapped OrcaSlicer profile. Click a slot to
+  open its edit dialog (see [Managing Filaments](#managing-filaments)).
+  If an ACE dryer is attached and active, a separate drying-status row
+  appears below the grid.
+
+Two banners can appear above the dashboard: an upload-ready banner when a
+GCode file finishes uploading (with Print / Assign Slots / Cancel actions),
+and a pause-reason banner when the printer pauses itself (e.g. filament
+runout).
+
+---
+
+## Printing
+
+### Uploading GCode
+
+Open the **Browser** tab (sidebar) → **Uploaded** sub-tab, and
+either drag a `.gcode`/`.bgcode` file onto the drop zone or click it to pick
+a file. Uploaded files get thumbnails (if embedded by the slicer), a search
+box, a status filter (All / Successful / Failed / New) and a sort order
+(date, name, print duration). Select the checkbox on a card to enter
+multi-select mode for bulk deletion.
+
+If you configured file-ready mode as "Print dialog" (see
+[Settings Reference](#settings-reference)), a dialog opens right after
+upload offering to start the print immediately or assign AMS slots first.
+With "banner" mode you instead get a persistent banner at
+the top of the screen with the same choices, so you can keep browsing
+before deciding.
+
+### Starting a print / assigning filament
+
+When starting a file that uses multiple filament channels (AMS/ACE slots),
+the **filament assignment dialog** opens automatically (or via "Assign
+Slots"). It lets you:
+
+- Map each GCode filament channel to a physical AMS/ACE slot, with a
+  mismatch warning if a channel's expected material/color doesn't match
+  what's actually loaded in the slot you pick.
+- Expand **✂ Skip objects** to deselect specific
+  printable objects (for multi-object plates) before the print starts —
+  the same object list and skip mechanism is also available mid-print from
+  the Progress card's "Objects" button.
+- Toggle **Auto-Leveling** for this print.
+- If Spoolman is configured, assign a specific spool to each slot right
+  from this dialog.
+
+Confirm with **▶ Print** to send the job to the printer.
+
+### Print-start behavior settings
+
+Under **Settings → Printer** you can control default behavior for
+every print:
+
+- **Default slot (single-color print)** —
+  which AMS slot to use automatically for single-material files.
+- **Auto-leveling before print** — run bed leveling before every print.
+- **Resonance compensation before print** — run input-shaper calibration
+  before every print.
+- **After upload: print-start behavior**
+  — dialog vs. banner, as described above.
+- **Turn camera on at print start** —
+  auto-start the camera stream whenever a print begins.
+- **Show warning for web-upload
+  prints** — an extra confirmation step for files uploaded through the
+  browser rather than sliced directly for this printer, to catch
+  wrong-printer-profile mistakes.
+
+### While printing
+
+The Progress card provides **Pause/Resume**, **Stop** (with a confirmation
+prompt), and **Objects** to skip specific objects on a multi-object plate
+mid-print.
+
+---
+
+## Managing Filaments
+
+### AMS / ACE slots
+
+Each slot tile on the Dashboard can be opened (click it) to edit:
+
+- **Color** — via a color picker, recent-color swatches, or "copy color
+  from slot" to match another slot.
+- **Material** — quick buttons for common materials, or free text.
+- **OrcaSlicer profile override** — pick a specific imported or built-in
+  OrcaSlicer filament profile for this slot. This is what gets sent to the
+  slicer during AMS sync instead of a generic "Generic PLA/PETG" fallback
+  (see the README's [OrcaSlicer-KX](README.md#-recommended-slicer) section
+  for why this matters).
+- A **feed** button to extrude/load filament for that slot directly from
+  the UI.
+
+If your printer has an ACE dryer unit, an additional drying panel appears
+below the AMS grid when drying is active, and slot edit dialogs let you
+configure drying **presets**: PLA, PLA+, PETG, TPU, ABS/ASA, PA/PC, and
+three freely-nameable Custom presets, each with its own temperature
+(30–80 °C) and remaining-time (h:m:s) setting. Presets can be edited and
+saved, or reset back to their defaults.
+
+### Importing your own OrcaSlicer profiles
+
+Under **Settings → Filament → OrcaSlicer-Profile** (or from the "★ Own
+profiles" link inside a slot's profile dropdown), open the import dialog
+and either drag a **ZIP** of your OrcaSlicer filament folder onto the drop
+zone, or upload individual **.json** profile files. In OrcaSlicer, that
+folder is reachable via **Help → Show Configuration Folder →
+user/<id>/filament/**. Imported profiles show up in every slot's profile
+dropdown under a "★ Own profiles" group and can be removed again from the
+same import dialog's list.
+
+### Filament-profile mapping and visible vendors
+
+Still under **Settings → Filament**:
+
+- **Filament profile mapping (per
+  slot)** — pin a fixed OrcaSlicer profile to each AMS slot so the bridge
+  always reports that profile during slicer sync, regardless of what
+  material/color is currently loaded.
+- **Visible vendors** — restrict which vendors show
+  up in the slot profile dropdown (useful if you only ever use a handful of
+  brands); leaving nothing selected shows all vendors. "Generic" and your
+  own imported profiles are always visible regardless of this filter.
+
+### Spoolman integration
+
+Configure the Spoolman server URL under **Settings → Integrations →
+Spoolman** (e.g. `http://spoolman:7912`) and a sync rate in seconds (`0`
+means "sync only when a print finishes"). Once connected, a
+**Spoolman — Slot assignment** panel appears under
+**Settings → Filament**, letting you assign a specific spool from your
+Spoolman inventory to each AMS slot. Filament usage is then tracked and
+reported to Spoolman automatically as you print, and the filament
+assignment dialog shown when starting a print also lets you pick/confirm
+spools per slot at print time.
+
+---
+
+## Multi-Printer Setup
+
+KX-Bridge can manage several printers from one running instance.
+
+- **Add a printer:** go to the **Printers** tab and click
+  **"+ Add printer"**. Enter the IP (name is
+  optional); credentials are fetched automatically, same as during first
+  setup. Each additional printer gets its own port (7126, 7127, …).
+- **Switch printers:** use the dropdown in the header (next to the printer
+  name), or open the **Printers** tab and click **"Switch"**
+  on any non-active printer's card. Each card also shows live status
+  (state, current file, progress bar, nozzle/bed temperature) fetched
+  directly from that printer's own bridge instance.
+- **Remove a printer:** click the **✕** button on its card in the
+  **Printers** tab; you'll be asked to confirm.
+
+---
+
+## Power Switch Feature
+
+The power switch feature lets the bridge turn an external **smart plug**
+on or off, and query its state — it is **not** a connection to the
+printer's own internal power management, since the printer has no
+remotely controllable power state of its own. You need a smart plug
+(commonly a Tasmota-flashed plug) wired between the wall outlet and the
+printer's power supply, reachable over HTTP from the bridge.
+
+Configure it under **Settings → Connection**, in the "Power Switch" card,
+with three
+URLs:
+
+- **Power-On URL** — called to switch the plug on.
+- **Power-Off URL** — called to switch the plug off.
+- **Status URL** — polled to show the current on/off state.
+
+For a Tasmota device, these are typically of the form:
+
+```
+http://192.168.x.x/cm?cmnd=Power%20on
+http://192.168.x.x/cm?cmnd=Power%20off
+http://192.168.x.x/cm?cmnd=Power
+```
+
+replacing `192.168.x.x` with the smart plug's own IP address (not the
+printer's). Once configured, a 🔌 power icon appears next to that
+printer's card in the **Printers** tab; click it to toggle the plug.
+Turning it off asks for confirmation, since it will cut power to whatever
+is plugged in — make sure nothing is printing first. The icon's color
+reflects the last known state (green = on, gray = off) as reported by the
+status URL.
+
+---
+
+## Settings Reference
+
+Settings are organized into tabs on the **Settings** panel:
+
+- **Connection** — printer name, printer IP, MQTT port,
+  MQTT username/password, device ID and mode ID (normally filled in
+  automatically by "Add printer"), plus the Power Switch URLs described
+  above.
+- **Printer** — default slot for single-color prints,
+  auto-leveling and resonance-compensation defaults, upload/print-start
+  behavior, camera auto-start, and the web-upload confirmation warning
+  (see [Printing](#printing)).
+- **Display** — UI language (DE/EN/ES/FR/IT/中文), light/dark
+  theme toggle, how often the bridge polls the printer for status updates,
+  and a verbose HTTP request logging toggle for troubleshooting.
+- **Filament** — OrcaSlicer profile import, per-slot profile mapping,
+  visible-vendor filtering, and (if Spoolman is connected) the Spoolman
+  slot-assignment panel — all described in
+  [Managing Filaments](#managing-filaments).
+- **Integrations** — Spoolman server URL and sync rate, and
+  an info box pointing to the `moonraker-obico.cfg` file used to configure
+  Obico (Obico itself is set up outside the bridge UI — see
+  [Camera / OrcaSlicer-KX / Obico](#camera--orcaslicer-kx--obico) below).
+- **System** — shows the current bridge version and lets you check for and
+  install updates directly from the browser, including a changelog preview.
+
+Most settings changes are applied via the **Save &
+Restart** button at the bottom of the Settings panel, which restarts the
+bridge process to apply them.
+
+---
+
+## Troubleshooting Basics
+
+- **Logs:** the **Console** tab shows a live event log with
+  filters by direction (RX/TX), level (errors/warnings), and topic (AMS,
+  print, info, status), plus a free-text filter and a download button for
+  the full log file.
+- **"Wrong MQTT credentials" on start:** re-add the printer via
+  "+ Add printer", or see the credential-refresh steps in the README's
+  [Troubleshooting](README.md#-troubleshooting) section.
+- **Printer not found / no LAN mode:** confirm LAN mode is enabled on the
+  printer's display and that the printer and bridge are on the same
+  network.
+- **Docker permission errors, upgrading from old versions, and other
+  install-level issues:** see the README's own
+  [Troubleshooting](README.md#-troubleshooting) section.
+
+For anything not covered here or in the README, please check or open an
+issue on the project's Gitea page:
+<https://gitea.it-drui.de/viewit/KX-Bridge-Release/issues>.
+
+---
+
+## Camera / OrcaSlicer-KX / Obico
+
+- **Camera:** the Dashboard's Camera tile plays the printer's live stream
+  directly; no separate setup is required beyond having the printer
+  connected.
+- **OrcaSlicer-KX:** for filament brand/color to sync correctly into
+  OrcaSlicer's AMS view, use the patched community build — see the
+  README's [Recommended Slicer](README.md#-recommended-slicer) section for
+  the download link and what it changes.
+- **Obico:** self-hosted failure-detection and time-lapse integration runs
+  through the separate `moonraker-obico` plugin/container, configured via
+  the config file referenced under **Settings → Integrations → Obico**.
+  Full setup instructions live in the README's
+  [Community & Integrations](README.md#-community--integrations) section.

NIGHTLY_CHANGELOG.md

@@ -1,6 +1,2 @@
 ## Changes in this build
 
-- Feat: `server/files/metadata` now uses the printer's own `buried/report` analytics event (fires once per print start, regardless of slicer) as a fallback for `size`/`estimated_time`/`layer_count` — fixes broken `size: 1`/`estimated_time: null` placeholders for files not in the bridge's own GCode store, e.g. prints started directly from Anycubic Slicer Next (Issue #102, thanks @fmontagna). Also surfaces the printer's storage usage (`storage_total_mb`/`storage_used_mb`) in `/api/state`.
-- Fix: **the bridge could get stuck in an endless reconnect loop after a printer disconnect, even once the printer was back online and reachable** — a container restart was the only way out. Two independent reconnect paths (the MQTT reader thread and the status-poll loop) could race into competing TLS handshakes, and the poll loop never noticed a dead MQTT session on its own since a failed send silently returned no data instead of raising an error. The bridge now reconnects automatically without manual intervention (Issue #105, thanks @p2l for the precise report).
-- Fix: the in-app update check on **stable** releases only ever looked at the single newest release on Gitea regardless of type — since nightly/dev prereleases publish far more often than stable ones, that newest release is almost always a prerelease, so the check found nothing and reported "no stable releases found" even though a newer stable release existed (Issue #104, thanks @Nerdinat0r).
-- Feat: printers with no MQTT-level standby/power-off command (i.e. all of them) can now be switched via an external smart plug (e.g. Tasmota) directly from the dashboard — configure a power-on/power-off/status URL per printer in Settings, and a power button with a live on/off indicator appears on that printer's card (Issue #103, thanks @ok24). Also fixes `config.ini` values containing a literal `%` (e.g. Tasmota's `cmnd=Power%20on` URLs) being rejected/corrupted by `configparser`'s default string interpolation.

README.md

@@ -33,6 +33,8 @@ officially tested or supported. Feedback welcome.</sub>
 >
 > 👉 Want to contribute? Please read [CONTRIBUTING.md](CONTRIBUTING.md) first.
 
+> 📖 Looking for day-to-day usage help? See the [User Manual](MANUAL.md). Building an integration? See the [API Reference](API.md).
+
 ---
 
 ## ✨ Features

config_loader.py

@@ -67,6 +67,7 @@ CONFIG_ENV_MAPPING = {
     "VIBRATION_COMPENSATION":   (CONFIG_SECTION_PRINT,   "vibration_compensation"),
     "CAMERA_ON_PRINT":          (CONFIG_SECTION_PRINT,   "camera_on_print"),
     "WEB_UPLOAD_WARNING":   (CONFIG_SECTION_PRINT,   "web_upload_warning"),
+    "DELETE_PRINTER_FILE_AFTER_PRINT": (CONFIG_SECTION_PRINT, "delete_printer_file_after_print"),
     "PRINT_START_DIALOG":   (CONFIG_SECTION_PRINT,   "print_start_dialog"),
     "BRIDGE_PRINTER_NAME":  (CONFIG_SECTION_BRIDGE,  "printer_name"),
     "BRIDGE_HOST_IP":       (CONFIG_SECTION_BRIDGE,  "host_ip"),
@@ -459,6 +460,7 @@ AUTO_LEVELING           = int(get("AUTO_LEVELING",          "1"))
 VIBRATION_COMPENSATION  = int(get("VIBRATION_COMPENSATION", "0"))
 CAMERA_ON_PRINT         = int(get("CAMERA_ON_PRINT",        "0"))
 WEB_UPLOAD_WARNING = int(get("WEB_UPLOAD_WARNING", "1"))
+DELETE_PRINTER_FILE_AFTER_PRINT = int(get("DELETE_PRINTER_FILE_AFTER_PRINT", "0"))
 PRINT_START_DIALOG = int(get("PRINT_START_DIALOG", get("FILE_READY_DIALOG", "1")))
 SPOOLMAN_SERVER    = get("SPOOLMAN_SERVER",    "")
 SPOOLMAN_SYNC_RATE = int(get("SPOOLMAN_SYNC_RATE", "0"))

env_loader.py

@@ -54,5 +54,6 @@ AUTO_LEVELING            = int(get("AUTO_LEVELING", "1"))
 VIBRATION_COMPENSATION   = int(get("VIBRATION_COMPENSATION", "0"))
 CAMERA_ON_PRINT          = int(get("CAMERA_ON_PRINT", "0"))
 WEB_UPLOAD_WARNING = int(get("WEB_UPLOAD_WARNING", "1"))
+DELETE_PRINTER_FILE_AFTER_PRINT = int(get("DELETE_PRINTER_FILE_AFTER_PRINT", "0"))
 PRINT_START_DIALOG = int(get("PRINT_START_DIALOG", get("FILE_READY_DIALOG", "1")))
 BRIDGE_HOST_IP   = get("BRIDGE_HOST_IP", "")

kobrax_client.py

@@ -101,6 +101,46 @@ def _parse_publish(pkt: bytes):
     return topic, payload
 
 
+def _enable_tcp_keepalive(sock: socket.socket) -> None:
+    """Without this, a printer that goes dark without a clean TCP close (e.g.
+    unplugged, not gracefully shut down) leaves the socket looking alive to
+    is_connected() for as long as the OS's default dead-connection timeout
+    (often 15+ minutes on Linux) - sendall() on a half-open connection is
+    buffered by the kernel and doesn't fail immediately, so the poll loop's
+    is_connected() check (kobrax_moonraker_bridge.py's _poll_loop) never
+    sees the failure it needs to flip kobra_state to "offline". Short
+    keepalive probes make the OS notice and fail the socket within seconds
+    instead. Linux/macOS only (TCP_KEEPIDLE/INTVL/CNT); best-effort on other
+    platforms - not fatal if unsupported.
+
+    SO_KEEPALIVE alone is NOT enough, verified live by unplugging a real
+    printer mid-connection: keepalive probes only fire while the connection
+    is idle (no unacknowledged data outstanding). If the printer disappears
+    while a send is still in flight - the common case, since the poll loop
+    sends a request roughly every poll_interval - the kernel instead retries
+    that specific send via the normal TCP retransmission timer
+    (tcp_retries2, default 15 attempts with exponential backoff = 13-30+
+    minutes on Linux), which keepalive settings don't affect at all.
+    TCP_USER_TIMEOUT (Linux-specific) closes that gap: it caps how long ANY
+    unacknowledged data may sit in the send queue before the kernel gives up
+    on the connection outright, regardless of which mechanism (keepalive or
+    retransmission) would otherwise still be retrying."""
+    try:
+        sock.setsockopt(socket.SOL_SOCKET, socket.SO_KEEPALIVE, 1)
+        if hasattr(socket, "TCP_KEEPIDLE"):
+            sock.setsockopt(socket.IPPROTO_TCP, socket.TCP_KEEPIDLE, 5)
+        elif hasattr(socket, "TCP_KEEPALIVE"):  # macOS
+            sock.setsockopt(socket.IPPROTO_TCP, socket.TCP_KEEPALIVE, 5)
+        if hasattr(socket, "TCP_KEEPINTVL"):
+            sock.setsockopt(socket.IPPROTO_TCP, socket.TCP_KEEPINTVL, 3)
+        if hasattr(socket, "TCP_KEEPCNT"):
+            sock.setsockopt(socket.IPPROTO_TCP, socket.TCP_KEEPCNT, 3)
+        if hasattr(socket, "TCP_USER_TIMEOUT"):
+            sock.setsockopt(socket.IPPROTO_TCP, socket.TCP_USER_TIMEOUT, 15000)
+    except OSError as e:
+        log.debug("TCP keepalive not fully supported on this platform: %s", e)
+
+
 # ---------------------------------------------------------------------------
 # KobraXClient
 # ---------------------------------------------------------------------------
@@ -189,6 +229,7 @@ class KobraXClient:
         # senders. Only the finished socket is swapped in under the lock (#53).
         _ai      = socket.getaddrinfo(self.host, self.port, socket.AF_INET, socket.SOCK_STREAM)
         raw      = socket.create_connection(_ai[0][4], timeout=5)
+        _enable_tcp_keepalive(raw)
         new_sock = ctx.wrap_socket(raw)
         log.info("TLS connected  cipher=%s", new_sock.cipher()[0])
 
@@ -252,19 +293,31 @@ class KobraXClient:
         with self._lock:
             return self._sock is not None
 
-    def _reconnect(self):
-        """Persistent reconnect: keeps retrying forever until the printer is
-        responds or disconnect() was called. Backoff caps at 60 s. The
-        first 5 attempts log as WARNING (acute connection issue), afterwards
-        only DEBUG to avoid log spam during long printer outages (e.g. switched
-        ausgeschaltet) zu vermeiden.
+    def _reconnect(self, wait_if_in_progress: bool = True, persist: bool = True):
+        """Reconnect the MQTT/TLS session. With persist=True (the default, used
+        by the reader-thread keepalive path) it keeps retrying forever until
+        the printer responds or disconnect() was called, backoff capped at 60s.
+        The first 5 attempts log as WARNING (acute connection issue), afterwards
+        only DEBUG to avoid log spam during long printer outages (e.g. switched off).
 
         Guarded by _reconnect_lock (Issue #105): if another thread's reconnect
-        is already in flight, this call waits for it to finish instead of
-        starting a second, competing _do_connect() - the printer likely only
+        is already in flight, this call normally waits for it to finish instead
+        of starting a second, competing _do_connect() - the printer likely only
         accepts one mTLS session at a time, so two parallel handshakes would
-        just interfere with each other and neither converges."""
+        just interfere with each other and neither converges.
+
+        wait_if_in_progress=False + persist=False are used by the poll loop's
+        publish()/publish_web(): that thread MUST return promptly so the poll
+        loop can observe the dead session (via is_connected()) and flip
+        kobra_state to "offline". It must neither block on the lock waiting for
+        the reader thread's persistent reconnect (wait_if_in_progress=False),
+        nor run the multi-minute backoff loop itself (persist=False -> at most
+        one immediate attempt). Otherwise the poll loop hangs inside publish()
+        for the entire outage and the dashboard stays stuck on the last known
+        state - the exact bug seen when a printer was unplugged mid-connection."""
         if not self._reconnect_lock.acquire(blocking=False):
+            if not wait_if_in_progress:
+                return self._sock is not None
             self._reconnect_lock.acquire()
             self._reconnect_lock.release()
             return self._sock is not None
@@ -290,6 +343,12 @@ class KobraXClient:
                     return True
                 except Exception as e:
                     attempt += 1
+                    if not persist:
+                        # One-shot: don't block the caller (poll loop) in the
+                        # backoff loop - leave persistent retrying to the
+                        # reader thread's keepalive path.
+                        log.debug("Reconnect (one-shot) failed: %s", e)
+                        return False
                     lvl = log.warning if attempt <= 5 else log.debug
                     lvl("Reconnect fehlgeschlagen (%s, Versuch %d), warte %ss…", e, attempt, delay)
                     # Split sleep so disconnect() breaks the loop faster.
@@ -522,7 +581,10 @@ class KobraXClient:
             self._pending_msgid.pop(msgid, None)
             if report_registered:
                 self._pending_report.pop(report_key, None)
-            if not self._reconnect():
+            # Non-blocking: never hang the poll-loop thread inside publish()
+            # while a reconnect is running / during backoff (see _reconnect
+            # docstring) - it must return so kobra_state can flip to "offline".
+            if not self._reconnect(wait_if_in_progress=False, persist=False):
                 return None
             # retry once after reconnect
             try:
@@ -569,10 +631,11 @@ class KobraXClient:
         except Exception as e:
             log.error("web send error: %s, reconnecting…", e)
             # Trigger a reconnect (like publish()); no retry because it is
-            # fire-and-forget - the next call will hit the fresh socket
-            # treffen.
+            # fire-and-forget - the next call will hit the fresh socket.
+            # Non-blocking for the same reason as publish() (see _reconnect
+            # docstring) - never hang this thread through a backoff loop.
             try:
-                self._reconnect()
+                self._reconnect(wait_if_in_progress=False, persist=False)
             except Exception:
                 pass
 

kobrax_moonraker_bridge.py

@@ -1073,6 +1073,11 @@ class KobraXBridge:
         self._store = store if store is not None else GCodeStore(args.data_dir)
         self._serve_dir_path: str = self._store._gcode_dir
         self._current_job_id: str = ""
+        # Filename of the file backing _current_job_id, kept alongside it so
+        # the "finished" handler can still delete it from the printer's own
+        # storage (Issue: delete-after-print) after self._state["filename"]
+        # has already been cleared as part of the terminal-state reset below.
+        self._current_job_filename: str = ""
         self._camera_autostarted: bool = False
         self._camera_user_stopped: bool = False  # user manually stopped the camera during a print
         self.camera_cache: CameraCache = CameraCache()
@@ -1396,6 +1401,7 @@ class KobraXBridge:
                         gcode_file_id=gf["id"],
                         printer_id=self._printer_id,
                     )
+                    self._current_job_filename = filename
                     log.info(f"Job started: {self._current_job_id} for {filename}")
             self._spoolman_slot_usage = {}
             self._spoolman_slot_reported = {}
@@ -1408,11 +1414,21 @@ class KobraXBridge:
             log.info(f"Job abgeschlossen: {self._current_job_id}")
             self._spoolman_notify_end()
             self._current_job_id = ""
+            # Optional cleanup (Settings -> Print): only for files that are
+            # also backed by the bridge's own GCode store - never for prints
+            # started directly from the printer/Anycubic Slicer, which would
+            # otherwise be deleted with no copy left anywhere (Issue: delete
+            # printer file after successful print). Deliberately only on a
+            # clean "finished" - stoped/canceled prints keep their file.
+            if getattr(self._args, "delete_printer_file_after_print", 0) and self._current_job_filename:
+                self._delete_printer_file_fire_and_forget(self._current_job_filename)
+            self._current_job_filename = ""
         elif kobra_state in ("stoped", "canceled") and self._current_job_id:
             self._store.finish_job(self._current_job_id, status="cancelled")
             log.info(f"Job abgebrochen: {self._current_job_id}")
             self._spoolman_notify_end()
             self._current_job_id = ""
+            self._current_job_filename = ""
 
         # Terminal states (successful finish AND stop/cancel) must leave the
         # same clean end state - a "finished" print used to only clear
@@ -1579,6 +1595,25 @@ class KobraXBridge:
         if payload.get("state") == "done" or payload.get("code") == 200:
             log.info(f"Skip response: state={payload.get('state')} code={payload.get('code')} skipped={skipped}")
 
+    def _delete_printer_file_fire_and_forget(self, filename: str) -> None:
+        """Deletes a file from the printer's own storage without waiting for
+        the response - called from _on_print(), which runs on the MQTT
+        reader thread itself, so blocking here (like _wait_for_file_action
+        does) would deadlock: the file/report reply that would unblock it is
+        dispatched from that same thread. Fire-and-forget is safe because the
+        bridge's own copy in the GCode store is what matters for correctness
+        here; a failed delete just leaves the printer's storage as it is
+        (Settings -> Print -> "Delete file from printer after successful print")."""
+        try:
+            self.client.publish(
+                "file", "deleteBatch",
+                {"root": "local", "files": [{"path": "/", "filename": filename}]},
+                timeout=0,
+            )
+            log.info(f"Requested printer-storage delete for {filename} after successful print")
+        except Exception as e:
+            log.warning(f"Delete-after-print request failed for {filename}: {e}")
+
     def _wait_for_file_action(self, action: str, send_fn, timeout: float = 8.0) -> dict | None:
         """Sends a file/* MQTT request (via send_fn, which must call
         self.client.publish(..., timeout=0) fire-and-forget) and blocks the
@@ -2243,12 +2278,32 @@ class KobraXBridge:
             return "", ""
         return vendor, family
 
-    def _match_profile_by_vendor_family(self, vendor: str, family: str) -> dict:
+    @staticmethod
+    def _rfid_variant_tokens(raw_type: str) -> list[str]:
+        """Tokens after "VENDOR TYPE" in a combined ACE-RFID string (e.g.
+        ["bas"] for "GEEETECH PLA Bas") - the truncated variant/serial that
+        distinguishes multiple profiles of the same (vendor, material family),
+        e.g. "Basic" vs. "Matte". Kept separate from _parse_combined_rfid_type()
+        so that function's 2-tuple signature (and its existing callers/tests)
+        stay unchanged (Issue #101)."""
+        tokens = raw_type.split()
+        return [t.lower() for t in tokens[2:]]
+
+    def _match_profile_by_vendor_family(self, vendor: str, family: str,
+                                         variant_tokens: list[str] | None = None) -> dict:
         """Find an imported/system filament profile by (vendor, material
         family) - used to auto-resolve a combined ACE-RFID type string to
         the user's already-imported OrcaSlicer profile (Issue #101), since
         the exact profile `name` never appears verbatim in the truncated
-        RFID string."""
+        RFID string.
+
+        When multiple profiles share the same (vendor, family) - e.g. "Geeetech
+        PLA Basic" and "Geeetech PLA Matte" both matching (Geeetech, PLA) -
+        variant_tokens (the RFID string's remaining tokens, e.g. ["bas"] for
+        "Basic") are scored against each candidate's name: a word-prefix match
+        scores higher than a plain substring match, so "bas" prefers "Basic"
+        over "Matte" or an unrelated profile name containing "bas" as noise.
+        Falls back to the first match when nothing disambiguates."""
         matches = [
             p for p in self._load_orca_filaments()
             if p.get("vendor", "").lower() == vendor.lower()
@@ -2256,12 +2311,28 @@ class KobraXBridge:
         ]
         if not matches:
             return {}
-        if len(matches) > 1:
-            log.debug(
-                f"_match_profile_by_vendor_family: {len(matches)} profiles match "
-                f"vendor={vendor!r} family={family!r}, using first: {matches[0].get('name')}"
-            )
-        return matches[0]
+        if len(matches) == 1 or not variant_tokens:
+            return matches[0]
+
+        best = matches[0]
+        best_score = -1
+        for p in matches:
+            name_words = p.get("name", "").lower().split()
+            score = 0
+            for tok in variant_tokens:
+                if any(w.startswith(tok) for w in name_words):
+                    score += 2
+                elif tok in p.get("name", "").lower():
+                    score += 1
+            if score > best_score:
+                best_score = score
+                best = p
+        log.debug(
+            f"_match_profile_by_vendor_family: {len(matches)} profiles match "
+            f"vendor={vendor!r} family={family!r}, variant_tokens={variant_tokens!r} "
+            f"-> {best.get('name')!r} (score={best_score})"
+        )
+        return best
 
     def _profile_material(self, profile: dict) -> str:
         """Material type (e.g. "PETG") of a saved slot profile, resolved by
@@ -2278,21 +2349,44 @@ class KobraXBridge:
 
     def _effective_slot_profile(self, global_idx: int, ams_material: str) -> dict:
         """Saved slot-profile override — but only while its material *family*
-        still matches the material currently loaded in the AMS.
+        still matches the material currently loaded in the AMS. Falls back to
+        auto-resolving a combined ACE-RFID type string (Issue #101) when there
+        is no (usable) manual override.
 
         Non-destructive suppression (Option A): when the family no longer matches
-        (e.g. a PETG profile but PLA loaded) we return {} → the slot falls back to
-        the generic default. The override stays in config.ini and reactivates as
-        soon as the matching material is loaded again. When the profile's family
-        is unknown we do NOT suppress (fail-safe)."""
+        (e.g. a PETG profile but PLA loaded) the override is skipped → falls
+        through to the RFID auto-match / generic default. The override stays in
+        config.ini and reactivates as soon as the matching material is loaded
+        again. When the profile's family is unknown we do NOT suppress (fail-safe).
+
+        Centralized here (rather than duplicated per caller) so every consumer -
+        the dashboard's /kx/filament/slots, Happy-Hare gate data, and the
+        OrcaSlicer lane-data sync - benefits from RFID auto-matching identically,
+        instead of only the one call site that happened to also call
+        _parse_combined_rfid_type() directly."""
+        # A combined ACE-RFID string ("GEEETECH PLA Bas") carries a vendor
+        # prefix that _material_family() alone can't see past (it only
+        # strips known polymer prefixes, so "GEEETECH PLA BAS" resolves to
+        # itself, not "PLA") - resolve the plain material family through the
+        # RFID parser first so the stale-profile guard below compares against
+        # the actual polymer family, not the raw combined string.
+        vendor, family = self._parse_combined_rfid_type(ams_material)
+        plain_material = family or ams_material
+
         profile = self._filament_profiles.get(global_idx) or {}
-        if not profile.get("name"):
-            return {}
-        prof_fam = self._material_family(self._profile_material(profile))
-        ams_fam  = self._material_family(ams_material)
-        if prof_fam and ams_fam and prof_fam != ams_fam:
-            return {}
-        return profile
+        if profile.get("name"):
+            prof_fam = self._material_family(self._profile_material(profile))
+            ams_fam  = self._material_family(plain_material)
+            if not (prof_fam and ams_fam and prof_fam != ams_fam):
+                return profile
+
+        if vendor:
+            variant_tokens = self._rfid_variant_tokens(ams_material)
+            auto = self._match_profile_by_vendor_family(vendor, family, variant_tokens)
+            if auto.get("name"):
+                return auto
+
+        return {}
 
     def _build_lane_data(self) -> dict:
       """Builds BBL AMS JSON for OrcaSlicer DevFilaSystemParser::ParseV1_0.
@@ -2338,30 +2432,19 @@ class KobraXBridge:
             # The vendor is sent along (tray_sub_brands + filament_vendor),
             # so a patched OrcaSlicer can match by brand + type +
             # color (analogous to SnapmakerPrinterAgent).
-            # Two-layer resolution for the filament hint sent to OrcaSlicer:
+            # Three-layer resolution for the filament hint sent to OrcaSlicer,
+            # all handled inside _effective_slot_profile() (Issue #101):
             #   1. User-Wahl (config.ini [filament_profiles]) — exakte Kontrolle
-            #   2. Generic fallback (_TRAY_INFO_IDX) per material type - no
+            #   2. Combined ACE-RFID "VENDOR TYPE SERIAL" string (e.g.
+            #      "GEEETECH PLA Bas") auto-matched against the user's
+            #      already-imported profile library. Not persisted to
+            #      config.ini - re-derives on every call, so a differently
+            #      tagged spool loaded later isn't stuck with a stale match.
+            #   3. Generic fallback (_TRAY_INFO_IDX) per material type - no
             #      vendor hint; OrcaSlicer then picks its own generic preset
-            # Stale-profile guard: only apply the override while its material
-            # family still matches the loaded filament (PETG profile + PLA
-            # loaded -> dropped).
             user_profile = self._effective_slot_profile(slot_index, material)
-            if not user_profile.get("name"):
-                # Third layer: auto-resolve a combined ACE-RFID "VENDOR TYPE
-                # SERIAL" string (e.g. "GEEETECH PLA Bas", from third-party
-                # RFID tools) against the user's already-imported profile
-                # library, instead of falling through to the neutral Generic
-                # fallback (Issue #101). Not persisted to config.ini - this
-                # re-derives on every _build_lane_data() call, so a
-                # differently-tagged spool loaded later isn't stuck with a
-                # stale match.
-                vendor_guess, family_guess = self._parse_combined_rfid_type(slot.get("type", ""))
-                if vendor_guess:
-                    auto_profile = self._match_profile_by_vendor_family(vendor_guess, family_guess)
-                    if auto_profile.get("name"):
-                        user_profile = auto_profile
-                        material = family_guess
             if user_profile.get("name"):
+                material = self._material_family(user_profile.get("type", material)) or material
                 vendor    = user_profile.get("vendor", "")
                 fila_name = user_profile.get("name", "")
                 tray_info_idx = user_profile.get("id") or self._TRAY_INFO_IDX.get(material, "OGFL99")
@@ -3547,6 +3630,7 @@ class KobraXBridge:
             printer_id=getattr(self._args, "device_id", "unknown"),
             filament_assignments=assignments,
         )
+        self._current_job_filename = filename
 
         return self._json_cors({"result": "ok", "filename": filename})
 
@@ -4996,6 +5080,7 @@ class KobraXBridge:
             "vibration_compensation": getattr(self._args, "vibration_compensation", 0),
             "camera_on_print":        getattr(self._args, "camera_on_print", 0),
             "web_upload_warning":     getattr(self._args, "web_upload_warning", 1),
+            "delete_printer_file_after_print": getattr(self._args, "delete_printer_file_after_print", 0),
             "print_start_dialog":     getattr(self._args, "print_start_dialog", 1),
             "poll_interval":    getattr(self._args, "poll_interval", 3),
             "verbose_http_log": getattr(self._args, "verbose_http_log", 0),
@@ -5037,6 +5122,7 @@ class KobraXBridge:
         cfg.set("print",      "vibration_compensation",  str(int(bool(data.get("vibration_compensation", getattr(self._args, "vibration_compensation", 0))))))
         cfg.set("print",      "camera_on_print",         str(int(bool(data.get("camera_on_print",        getattr(self._args, "camera_on_print",        0))))))
         cfg.set("print",      "web_upload_warning", str(int(bool(data.get("web_upload_warning", getattr(self._args, "web_upload_warning", 1))))))
+        cfg.set("print",      "delete_printer_file_after_print", str(int(bool(data.get("delete_printer_file_after_print", getattr(self._args, "delete_printer_file_after_print", 0))))))
         cfg.set("print",      "print_start_dialog", str(int(bool(data.get("print_start_dialog", getattr(self._args, "print_start_dialog", 1))))))
         if "poll_interval" in data:
             try:
@@ -6219,6 +6305,10 @@ def main():
     parser.add_argument("--vibration-compensation",  type=int, default=env_loader.VIBRATION_COMPENSATION)
     parser.add_argument("--camera-on-print",         type=int, default=env_loader.CAMERA_ON_PRINT)
     parser.add_argument("--web-upload-warning", type=int, default=env_loader.WEB_UPLOAD_WARNING)
+    parser.add_argument("--delete-printer-file-after-print", type=int,
+                        default=env_loader.DELETE_PRINTER_FILE_AFTER_PRINT,
+                        help="After a successful print, delete the file from the printer's "
+                             "own storage if it's also in the bridge's own GCode store")
     parser.add_argument("--print-start-dialog", dest="print_start_dialog", type=int, default=env_loader.PRINT_START_DIALOG)
     parser.add_argument("--file-ready-dialog",  dest="print_start_dialog", type=int)
     parser.add_argument("--spoolman-server",    default=env_loader.SPOOLMAN_SERVER,

tests/test_ace_rfid_vendor_matching.py

@@ -109,3 +109,128 @@ def test_build_lane_data_plain_type_still_uses_generic_fallback():
     tray = lane["ams"][0]["tray"][0]
     assert tray["name"] == "Generic PLA"
     assert tray["vendor_name"] == "Generic"
+
+
+# ─── Centralized matching via _effective_slot_profile() ────────────────────
+#
+# The bug reported in Issue #101 by @Blaim (nightly45 not working, despite
+# _parse_combined_rfid_type()/_match_profile_by_vendor_family() existing):
+# those two helpers were ONLY ever invoked inside _build_lane_data(), which
+# is only reached when OrcaSlicer polls the Moonraker lane_data endpoint -
+# never as part of the real MQTT receive path (_on_multicolor_box ->
+# self._ams_slots -> _push_status_update -> dashboard / /kx/filament/slots).
+# So the dashboard and the Happy-Hare gate data never saw a match, matching
+# exactly what the user's screenshots showed. Fixed by moving the matching
+# logic into _effective_slot_profile() itself, which all three consumers
+# already call.
+
+def test_effective_slot_profile_auto_resolves_raw_rfid_string_without_override():
+    """The core Issue #101 regression: _effective_slot_profile() itself (not
+    just _build_lane_data()) must resolve a combined RFID string when there is
+    no manual per-slot override in config.ini."""
+    b = _bridge()
+    b._filament_profiles = {}
+    profile = b._effective_slot_profile(0, "GEEETECH PLA Bas")
+    assert profile.get("name") == "Geeetech PLA Basic"
+    assert profile.get("vendor") == "Geeetech"
+
+
+def test_effective_slot_profile_manual_override_still_wins():
+    b = _bridge()
+    b._filament_profiles = {0: {"id": "OGFL99", "name": "Generic PLA", "vendor": "Generic"}}
+    profile = b._effective_slot_profile(0, "GEEETECH PLA Bas")
+    assert profile.get("name") == "Generic PLA"
+
+
+def test_effective_slot_profile_plain_type_no_override_returns_empty():
+    """Regression guard: a plain type="PLA" slot with no override must still
+    fall through to {} (the caller's own generic-name fallback), not be
+    treated as an RFID string."""
+    b = _bridge()
+    b._filament_profiles = {}
+    assert b._effective_slot_profile(0, "PLA") == {}
+
+
+def _multicolor_box_report(raw_type: str, color=(238, 190, 152)) -> dict:
+    """A realistic multiColorBox/report payload for one ACE box (id=0) with
+    a toolhead (id=-1), matching the real Kobra X topology captured live
+    during Issue #100/#101 investigation - drives _detect_filament_mode()
+    to "ace_hub", same as on real hardware."""
+    return {
+        "state": "success",
+        "data": {
+            "head_tools_model": 1,
+            "multi_color_box": [
+                {
+                    "id": -1, "loaded_slot": -1,
+                    "slots": [{"index": 0, "status": 0, "type": "", "color": [0, 0, 0]}],
+                },
+                {
+                    "id": 0, "loaded_slot": -1,
+                    "slots": [
+                        {"index": 0, "status": 0, "type": "", "color": [0, 0, 0]},
+                        {"index": 1, "status": 0, "type": "", "color": [0, 0, 0]},
+                        {
+                            "index": 2, "status": 5, "type": raw_type,
+                            "color": list(color), "sku": "",
+                        },
+                        {"index": 3, "status": 0, "type": "", "color": [0, 0, 0]},
+                    ],
+                },
+            ],
+        },
+    }
+
+
+def test_on_multicolor_box_end_to_end_resolves_rfid_slot_for_dashboard():
+    """End-to-end regression test for Issue #101: feed a raw MQTT
+    multiColorBox/report payload through the real receive path
+    (_on_multicolor_box), then check that the dashboard-facing
+    /kx/filament/slots data (handle_kx_filament_slots) - which is what
+    populates the dashboard's window._slotProfileMap in the browser -
+    actually reflects the matched profile, not the raw "GEEETECH PLA Bas"
+    string. This is the exact path that was broken and untested before."""
+    b = _bridge()
+    b._filament_profiles = {}
+    b._on_multicolor_box(_multicolor_box_report("GEEETECH PLA Bas"))
+
+    assert b._filament_mode == "ace_hub"
+    # global_index 6 = box_id 0 * 4 + local slot 2 in ace_hub mode's ACE block
+    slot = next(s for s in b._ams_slots if s.get("type") == "GEEETECH PLA Bas")
+    global_idx = slot["global_index"]
+
+    profile = b._effective_slot_profile(global_idx, slot["type"])
+    assert profile.get("name") == "Geeetech PLA Basic"
+    assert profile.get("vendor") == "Geeetech"
+
+
+def test_match_profile_by_vendor_family_disambiguates_via_variant_tokens():
+    """Issue #101 follow-up (@Blaim): two profiles of the same vendor+family
+    ("Geeetech PLA Basic" vs. "Geeetech PLA Matte") must resolve to the one
+    matching the RFID string's truncated variant token ("bas" -> Basic),
+    not just "whichever loads first"."""
+    profiles = USER_PROFILES + [
+        {"id": "GTPLA03", "name": "Geeetech PLA Matte", "vendor": "Geeetech", "type": "PLA", "color": "", "is_user": True},
+    ]
+    b = _bridge(profiles)
+
+    basic = b._match_profile_by_vendor_family("Geeetech", "PLA", ["bas"])
+    assert basic.get("name") == "Geeetech PLA Basic"
+
+    matte = b._match_profile_by_vendor_family("Geeetech", "PLA", ["mat"])
+    assert matte.get("name") == "Geeetech PLA Matte"
+
+
+def test_match_profile_by_vendor_family_no_variant_tokens_falls_back_to_first():
+    profiles = USER_PROFILES + [
+        {"id": "GTPLA03", "name": "Geeetech PLA Matte", "vendor": "Geeetech", "type": "PLA", "color": "", "is_user": True},
+    ]
+    b = _bridge(profiles)
+    profile = b._match_profile_by_vendor_family("Geeetech", "PLA")
+    assert profile.get("name") == "Geeetech PLA Basic"
+
+
+def test_rfid_variant_tokens_extracts_tokens_after_vendor_and_family():
+    assert KobraXBridge._rfid_variant_tokens("GEEETECH PLA Bas") == ["bas"]
+    assert KobraXBridge._rfid_variant_tokens("GEEETECH PLA") == []
+    assert KobraXBridge._rfid_variant_tokens("PLA") == []

tests/test_delete_printer_file_after_print.py

@@ -0,0 +1,157 @@
+"""Optional auto-delete of a printed file from the printer's own storage
+after a successful print (Settings -> Print -> "Delete file from printer
+after successful print").
+
+Only applies to files that are also backed by the bridge's own GCode store
+(otherwise the file would be gone with no copy left anywhere) and only on a
+clean "finished" state - not on stoped/canceled prints, and never when the
+setting is off (the default).
+"""
+import argparse
+import tempfile
+from unittest.mock import MagicMock
+
+from kobrax_moonraker_bridge import GCodeStore, KobraXBridge
+
+
+def _bridge(delete_after_print=1):
+    c = MagicMock()
+    c.callbacks = {}
+    c.connected = False
+    args = argparse.Namespace(
+        printer_ip="", mqtt_port=9883, username="", password="",
+        mode_id="20030", device_id="", host="127.0.0.1", port=7125,
+        data_dir=tempfile.mkdtemp(prefix="kxdelafterprint-"),
+        delete_printer_file_after_print=delete_after_print,
+    )
+    store = GCodeStore(args.data_dir)
+    b = KobraXBridge(c, args=args, store=store)
+    return b, c
+
+
+def _seed_file(bridge, filename="test.gcode"):
+    file_id = "abc123"
+    bridge._store.save_file(file_id, filename, b"; gcode content")
+    return file_id
+
+
+def _print_report(state, filename=None):
+    payload = {"state": state, "data": {}}
+    if filename is not None:
+        payload["data"]["filename"] = filename
+    return payload
+
+
+def test_finished_print_deletes_printer_file_when_enabled_and_in_store():
+    b, c = _bridge(delete_after_print=1)
+    _seed_file(b, "test.gcode")
+
+    b._on_print(_print_report("printing", "test.gcode"))
+    assert b._current_job_id
+    assert b._current_job_filename == "test.gcode"
+
+    b._on_print(_print_report("finished"))
+
+    delete_calls = [
+        call for call in c.publish.call_args_list
+        if call.args[:2] == ("file", "deleteBatch")
+    ]
+    assert len(delete_calls) == 1
+    payload = delete_calls[0].args[2]
+    assert payload == {"root": "local", "files": [{"path": "/", "filename": "test.gcode"}]}
+
+
+def test_finished_print_no_delete_when_setting_disabled():
+    b, c = _bridge(delete_after_print=0)
+    _seed_file(b, "test.gcode")
+
+    b._on_print(_print_report("printing", "test.gcode"))
+    b._on_print(_print_report("finished"))
+
+    delete_calls = [
+        call for call in c.publish.call_args_list
+        if call.args[:2] == ("file", "deleteBatch")
+    ]
+    assert delete_calls == []
+
+
+def test_finished_print_no_delete_when_file_not_in_bridge_store():
+    """Files started directly from the printer/Anycubic Slicer aren't in the
+    bridge's own GCode store - must never be deleted, since that would leave
+    no copy anywhere."""
+    b, c = _bridge(delete_after_print=1)
+    # No _seed_file() call - the file is not in the store.
+
+    b._on_print(_print_report("printing", "not_in_store.gcode"))
+    assert not b._current_job_id  # no store match -> no job tracked either
+
+    b._on_print(_print_report("finished"))
+
+    delete_calls = [
+        call for call in c.publish.call_args_list
+        if call.args[:2] == ("file", "deleteBatch")
+    ]
+    assert delete_calls == []
+
+
+def test_canceled_print_does_not_delete_file():
+    """Only a clean "finished" triggers the delete - a stopped/canceled
+    print keeps its file, since the user may want to retry it."""
+    b, c = _bridge(delete_after_print=1)
+    _seed_file(b, "test.gcode")
+
+    b._on_print(_print_report("printing", "test.gcode"))
+    b._on_print(_print_report("canceled"))
+
+    delete_calls = [
+        call for call in c.publish.call_args_list
+        if call.args[:2] == ("file", "deleteBatch")
+    ]
+    assert delete_calls == []
+    assert b._current_job_filename == ""
+
+
+def test_stoped_print_does_not_delete_file():
+    b, c = _bridge(delete_after_print=1)
+    _seed_file(b, "test.gcode")
+
+    b._on_print(_print_report("printing", "test.gcode"))
+    b._on_print(_print_report("stoped"))
+
+    delete_calls = [
+        call for call in c.publish.call_args_list
+        if call.args[:2] == ("file", "deleteBatch")
+    ]
+    assert delete_calls == []
+
+
+def test_current_job_filename_reset_after_finished():
+    """Regression guard: _current_job_filename must not leak into the next
+    print's finished-handling if that next print isn't itself tracked."""
+    b, c = _bridge(delete_after_print=1)
+    _seed_file(b, "test.gcode")
+
+    b._on_print(_print_report("printing", "test.gcode"))
+    b._on_print(_print_report("finished"))
+    assert b._current_job_filename == ""
+
+    # A second "finished" with no new job in between must not re-trigger a delete.
+    c.publish.reset_mock()
+    b._on_print(_print_report("finished"))
+    delete_calls = [
+        call for call in c.publish.call_args_list
+        if call.args[:2] == ("file", "deleteBatch")
+    ]
+    assert delete_calls == []
+
+
+def test_delete_publish_failure_does_not_raise():
+    """A broken MQTT send during the delete request must not propagate out
+    of _on_print() - it runs on the MQTT reader thread, and an unhandled
+    exception there would break processing of subsequent messages."""
+    b, c = _bridge(delete_after_print=1)
+    _seed_file(b, "test.gcode")
+    c.publish.side_effect = RuntimeError("send failed")
+
+    b._on_print(_print_report("printing", "test.gcode"))
+    b._on_print(_print_report("finished"))  # must not raise

tests/test_mqtt_reconnect.py

@@ -92,3 +92,61 @@ def test_reconnect_second_waiter_returns_after_first_completes():
 
     assert result is True
     assert c._sock is not None
+
+
+def test_reconnect_non_blocking_returns_immediately_while_reconnect_in_progress():
+    """The poll-loop path (publish/publish_web) must NOT block while the reader
+    thread's persistent reconnect is running its multi-minute backoff loop -
+    it has to return so the poll loop can flip kobra_state to "offline".
+    A printer unplugged mid-connection otherwise left the dashboard stuck on
+    the last known state indefinitely (Issue #103 follow-up)."""
+    c = _client()
+    c._running = True
+    started = threading.Event()
+    release = threading.Event()
+
+    def slow_persistent_do_connect():
+        started.set()
+        # Simulate the printer still being gone: never succeeds until released.
+        release.wait(timeout=5.0)
+        raise OSError("still unreachable")
+
+    c._do_connect = slow_persistent_do_connect
+
+    # First reconnect (reader-thread style): persistent, holds the lock, stuck
+    # in backoff.
+    t1 = threading.Thread(target=lambda: c._reconnect(persist=True), daemon=True)
+    t1.start()
+    assert started.wait(timeout=2.0)
+
+    # Poll-loop style call must return basically instantly, not block on t1.
+    t0 = time.time()
+    result = c._reconnect(wait_if_in_progress=False, persist=False)
+    elapsed = time.time() - t0
+
+    assert elapsed < 0.5, f"non-blocking reconnect blocked for {elapsed:.2f}s"
+    assert result is False  # socket is down while the other reconnect churns
+
+    release.set()  # let the daemon thread unwind
+
+
+def test_reconnect_one_shot_does_not_loop_on_failure():
+    """persist=False must attempt the handshake at most once and return,
+    instead of entering the backoff loop (which would block the caller)."""
+    c = _client()
+    c._running = True
+    attempts = []
+
+    def failing_do_connect():
+        attempts.append(1)
+        raise OSError("unreachable")
+
+    c._do_connect = failing_do_connect
+
+    t0 = time.time()
+    result = c._reconnect(persist=False)
+    elapsed = time.time() - t0
+
+    assert result is False
+    assert len(attempts) == 1  # exactly one attempt, no backoff retries
+    assert elapsed < 0.5

tests/test_tcp_keepalive.py

@@ -0,0 +1,80 @@
+"""TCP keepalive + TCP_USER_TIMEOUT on the MQTT socket (Issue #103 follow-up).
+
+Without these, a printer that disappears without a clean TCP close
+(unplugged, not gracefully shut down) leaves the socket looking alive to
+is_connected() for as long as the OS's default dead-connection timeout -
+often 15+ minutes on Linux - since sendall() on a half-open connection is
+buffered by the kernel and doesn't fail immediately. This left the
+dashboard's printer-state indicator stuck showing the last known state
+(e.g. green "ready") long after the printer was actually unreachable,
+reported when testing the smart-plug power-switch feature by physically
+unplugging the printer.
+
+Verified live (real printer, physically unplugged) that SO_KEEPALIVE alone
+is not sufficient: keepalive probes only fire on an idle connection, but if
+the printer disappears while a send is still unacknowledged - the normal
+case, since the poll loop is sending every few seconds - the kernel's
+regular TCP retransmission timer takes over instead (tcp_retries2, 13-30+
+minutes on Linux), which keepalive settings don't affect. TCP_USER_TIMEOUT
+closes that gap by capping how long ANY unacknowledged data may sit in the
+send queue, regardless of which retry mechanism would otherwise still be
+running.
+"""
+import socket
+from unittest.mock import MagicMock
+
+import pytest
+
+from kobrax_client import _enable_tcp_keepalive
+
+
+def test_enable_tcp_keepalive_sets_so_keepalive():
+    s = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
+    try:
+        _enable_tcp_keepalive(s)
+        assert s.getsockopt(socket.SOL_SOCKET, socket.SO_KEEPALIVE) == 1
+    finally:
+        s.close()
+
+
+@pytest.mark.skipif(not hasattr(socket, "TCP_KEEPIDLE"), reason="Linux-specific option")
+def test_enable_tcp_keepalive_sets_short_idle_and_interval():
+    s = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
+    try:
+        _enable_tcp_keepalive(s)
+        idle = s.getsockopt(socket.IPPROTO_TCP, socket.TCP_KEEPIDLE)
+        intvl = s.getsockopt(socket.IPPROTO_TCP, socket.TCP_KEEPINTVL)
+        cnt = s.getsockopt(socket.IPPROTO_TCP, socket.TCP_KEEPCNT)
+        # Short enough that a dead connection is detected within a couple of
+        # poll cycles (default poll_interval is 3s), not the OS default of
+        # minutes.
+        assert idle <= 10
+        assert intvl <= 5
+        assert cnt <= 5
+    finally:
+        s.close()
+
+
+@pytest.mark.skipif(not hasattr(socket, "TCP_USER_TIMEOUT"), reason="Linux-specific option")
+def test_enable_tcp_keepalive_sets_user_timeout():
+    """The critical fix, verified live against a real printer: without this,
+    a dead connection with unacknowledged data in flight is only detected
+    after the OS's normal TCP retransmission timeout (13-30+ minutes on
+    Linux), not the keepalive interval."""
+    s = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
+    try:
+        _enable_tcp_keepalive(s)
+        user_timeout_ms = s.getsockopt(socket.IPPROTO_TCP, socket.TCP_USER_TIMEOUT)
+        # Short enough that a dead connection with in-flight data is detected
+        # within a couple of poll cycles, not tens of minutes.
+        assert 0 < user_timeout_ms <= 20000
+    finally:
+        s.close()
+
+
+def test_enable_tcp_keepalive_does_not_raise_on_unsupported_platform():
+    """A platform without TCP_KEEPIDLE/INTVL/CNT (e.g. some Windows builds)
+    must not crash the connection attempt - keepalive is best-effort."""
+    s = MagicMock()
+    s.setsockopt.side_effect = OSError("unsupported")
+    _enable_tcp_keepalive(s)  # must not raise

web/themes/default/app.js

@@ -464,6 +464,8 @@ function applyLang(){
   setText('opt-file-ready-banner',T.settings_file_ready_banner);
   setText('lbl-camera-on-print',T.settings_camera_on_print);
   setText('lbl-web-upload-warning',T.settings_web_upload_warning);
+  setText('lbl-delete-printer-file-after-print',T.settings_delete_printer_file_after_print||'Delete file from printer after successful print');
+  setText('lbl-delete-printer-file-after-print-hint',T.settings_delete_printer_file_after_print_hint||'Only applies to prints started through this bridge (files it uploaded itself) - prints started directly from the printer or Anycubic Slicer are never deleted, since no copy of those exists anywhere else.');
   setText('fd-options-title',T.fd_options_title);
   setText('fd-lbl-auto-leveling',T.print_auto_leveling);
 
@@ -1141,6 +1143,7 @@ function openSettings(){
     var cop=document.getElementById('s-camera-on-print');if(cop)cop.checked=!!d.camera_on_print;
     var frm=document.getElementById('s-file-ready-mode');if(frm)frm.value=(d.print_start_dialog===undefined?'1':String(d.print_start_dialog?1:0));
     var wuw=document.getElementById('s-web-upload-warning');if(wuw)wuw.checked=(d.web_upload_warning===undefined?true:!!d.web_upload_warning);
+    var dpfap=document.getElementById('s-delete-printer-file-after-print');if(dpfap)dpfap.checked=!!d.delete_printer_file_after_print;
     // Poll-Intervall (Sekunden) — Backend hat Vorrang vor localStorage
     var pi=document.getElementById('s-poll-interval');
     if(pi){
@@ -1912,6 +1915,7 @@ function saveSettings(){
     camera_on_print:         (document.getElementById('s-camera-on-print')||{}).checked?1:0,
     print_start_dialog: parseInt((document.getElementById('s-file-ready-mode')||{}).value||'1',10),
     web_upload_warning:webUploadWarning,
+    delete_printer_file_after_print: (document.getElementById('s-delete-printer-file-after-print')||{}).checked?1:0,
     poll_interval:    Math.min(60,Math.max(1,parseInt((document.getElementById('s-poll-interval')||{}).value,10)||3)),
     verbose_http_log: (document.getElementById('s-verbose-http-log')||{}).checked?1:0,
     spoolman_server:  (document.getElementById('s-spoolman-url')||{}).value||'',

web/themes/default/index.html

@@ -603,6 +603,11 @@
                 <input type="checkbox" id="s-web-upload-warning" style="width:auto;margin:0">
                 <label id="lbl-web-upload-warning" style="margin:0;cursor:pointer" for="s-web-upload-warning">Warnung bei Web-Upload-Druck anzeigen</label>
               </div>
+              <div class="modal-field" style="flex-direction:row;align-items:center;gap:10px">
+                <input type="checkbox" id="s-delete-printer-file-after-print" style="width:auto;margin:0">
+                <label id="lbl-delete-printer-file-after-print" style="margin:0;cursor:pointer" for="s-delete-printer-file-after-print">Delete file from printer after successful print</label>
+              </div>
+              <small id="lbl-delete-printer-file-after-print-hint" style="color:var(--txt2)"></small>
             </div>
           </div>
 

web/translations/de.json

@@ -293,6 +293,8 @@
   "settings_visible_vendors_save": "Auswahl speichern",
   "settings_visible_vendors_save_label": "Auswahl speichern",
   "settings_web_upload_warning": "Warnung bei Web-Upload-Druck anzeigen",
+  "settings_delete_printer_file_after_print": "Datei nach erfolgreichem Druck vom Drucker löschen",
+  "settings_delete_printer_file_after_print_hint": "Gilt nur für Drucke, die über diese Bridge gestartet wurden (selbst hochgeladene Dateien) - direkt am Drucker oder über Anycubic Slicer gestartete Drucke werden nie gelöscht, da davon sonst keine Kopie mehr existiert.",
   "sf_all": "Alle",
   "sf_err": "✗ Fehler",
   "sf_new": "Neu",

web/translations/en.json

@@ -293,6 +293,8 @@
   "settings_visible_vendors_save": "Save selection",
   "settings_visible_vendors_save_label": "Save selection",
   "settings_web_upload_warning": "Show warning when printing web uploads",
+  "settings_delete_printer_file_after_print": "Delete file from printer after successful print",
+  "settings_delete_printer_file_after_print_hint": "Only applies to prints started through this bridge (files it uploaded itself) - prints started directly from the printer or Anycubic Slicer are never deleted, since no copy of those exists anywhere else.",
   "sf_all": "All",
   "sf_err": "✗ Failed",
   "sf_new": "New",

web/translations/es.json

@@ -293,6 +293,8 @@
   "settings_visible_vendors_save": "Guardar selección",
   "settings_visible_vendors_save_label": "Guardar selección",
   "settings_web_upload_warning": "Mostrar advertencia al imprimir subidas web",
+  "settings_delete_printer_file_after_print": "Eliminar archivo de la impresora tras una impresión exitosa",
+  "settings_delete_printer_file_after_print_hint": "Solo aplica a impresiones iniciadas a través de este bridge (archivos que él mismo subió) - las impresiones iniciadas directamente desde la impresora o Anycubic Slicer nunca se eliminan, ya que no existe ninguna copia en otro lugar.",
   "sf_all": "Todos",
   "sf_err": "✗ Fallido",
   "sf_new": "Nuevo",

web/translations/fr.json

@@ -279,6 +279,8 @@
   "settings_visible_vendors_save": "Enregistrer la sélection",
   "settings_visible_vendors_save_label": "Enregistrer la sélection",
   "settings_web_upload_warning": "Afficher un avertissement lors de l'impression de fichiers web",
+  "settings_delete_printer_file_after_print": "Supprimer le fichier de l'imprimante après une impression réussie",
+  "settings_delete_printer_file_after_print_hint": "S'applique uniquement aux impressions lancées via ce bridge (fichiers qu'il a lui-même téléversés) - les impressions lancées directement depuis l'imprimante ou Anycubic Slicer ne sont jamais supprimées, car aucune copie n'existe ailleurs.",
   "sf_all": "Tout",
   "sf_err": "✗ Échoués",
   "sf_new": "Nouveau",

web/translations/it.json

@@ -279,6 +279,8 @@
   "settings_visible_vendors_save": "Salva selezione",
   "settings_visible_vendors_save_label": "Salva selezione",
   "settings_web_upload_warning": "Mostra un avviso quando si stampano caricamenti web",
+  "settings_delete_printer_file_after_print": "Elimina il file dalla stampante dopo una stampa riuscita",
+  "settings_delete_printer_file_after_print_hint": "Si applica solo alle stampe avviate tramite questo bridge (file caricati da esso) - le stampe avviate direttamente dalla stampante o da Anycubic Slicer non vengono mai eliminate, poiché non ne esiste alcuna copia altrove.",
   "sf_all": "Tutti",
   "sf_err": "✗ Fallito",
   "sf_new": "Nuovo",

web/translations/zh-cn.json

@@ -293,6 +293,8 @@
   "settings_visible_vendors_save": "保存选择",
   "settings_visible_vendors_save_label": "保存选择",
   "settings_web_upload_warning": "打印网页上传文件时显示警告",
+  "settings_delete_printer_file_after_print": "打印成功后从打印机删除文件",
+  "settings_delete_printer_file_after_print_hint": "仅适用于通过此网桥启动的打印（即由网桥自己上传的文件）——直接从打印机或 Anycubic Slicer 启动的打印任务永远不会被删除，因为它们没有其他备份。",
   "sf_all": "全部",
   "sf_err": "✗ 失败",
   "sf_new": "新",