diff options
| author | Luke Hoersten <[email protected]> | 2026-06-13 21:31:28 -0500 |
|---|---|---|
| committer | Luke Hoersten <[email protected]> | 2026-06-13 21:31:28 -0500 |
| commit | f5b9f2f81fa71361418a475d801eb4adefcbc027 (patch) | |
| tree | 7f0f35e9c9be5895cf71a03bd24532596e94879f /README.md | |
| parent | 072504c2dfa19e4311e16e5c445d9cc19680dcf5 (diff) | |
Reframe protocol as REST peers; mDNS-SD discovery; explicit no-ack
Three structural shifts:
1. Discovery section spells out mDNS-SD. The ESP32 runs the ESP-IDF
mdns responder and serves all .local records itself — A, SRV, PTR,
TXT. Scrypted browses _scrypted-viewport._tcp.local with a Node
mDNS-SD library that hits the multicast layer directly. Scrypted
uses the IP from the SRV/A record, NOT OS-level .local hostname
resolution. Re-browse every few minutes for DHCP renumbering.
Manual host:port fallback for non-mDNS deployments.
2. Drop "callback" framing. The device and Scrypted are HTTP peers,
both exposing POST /state with the same body shape {viewport,state}.
Either side can push to the other to set state. Section is renamed
"Device -> Scrypted POST /state". No new vocabulary or semantics —
just the truth about what the protocol is.
- state_post_failures (was: callback_failures)
- state_client.c (was: callback_client.c)
- "callback" replaced with "outbound POST" / "inbound POST" / just
"POST /state" throughout.
3. Explicit no-application-level-ack clause. HTTP 2xx is transport-
only. Device does not retry, does not block subsequent state
changes on the response, does not treat 5xx as anything beyond a
counter increment. Idempotency + /frame 409 + each side's idle
timer cover every failure mode. Scrypted-side logic must not wait
for the device to confirm a state change.
Also: Scrypted Integration code in v2 guide rewritten to discover via
bonjour-service browse rather than hardcoded viewport-<name>.local
URLs. urlFor(name) function resolves the discovered IP per call.
mdns_service.c module spec now explicitly notes Scrypted uses the
browse, not the hostname.
Diffstat (limited to 'README.md')
| -rw-r--r-- | README.md | 217 |
1 files changed, 161 insertions, 56 deletions
@@ -17,7 +17,7 @@ Design goals: Scrypted owns rendering, overlays, camera selection and interaction logic. -Scrypted Viewport owns Ethernet, JPEG decode, display, touch input and callback delivery. +Scrypted Viewport owns Ethernet, JPEG decode, display, touch input and outbound state-change POSTs. ## Hardware @@ -50,17 +50,31 @@ Scrypted must render JPEGs at the **effective** resolution. The device does not HTTP listen port: `80`. -mDNS service: `_scrypted-viewport._tcp.local` advertised on port 80. +Trust model: LAN-only, no auth, no TLS. Deploy on a trusted VLAN. -mDNS TXT records: -- `version=1.0.0` -- `resolution=<effective>` (e.g. `480x800` for portrait, `800x480` for landscape) -- `orientation=<portrait|landscape>` -- `name=<display name>` (empty until `/config`) +## Discovery -Hostname: `viewport.local` before configuration, `viewport-<display>.local` after. +The ESP32 publishes itself via **mDNS-SD (service discovery)**. Scrypted discovers viewports by browsing the service; it does not need OS-level `.local` hostname resolution. -Trust model: LAN-only, no auth, no TLS. Deploy on a trusted VLAN. +The device runs an mDNS responder (ESP-IDF `mdns` component) that serves all of the following from itself — no external DNS server is involved: + +- **Hostname / A record**: `viewport.local` before configuration, `viewport-<name>.local` after (e.g. `viewport-mudroom.local`). The `viewport-` prefix is a namespace that avoids collisions with other LAN devices. +- **Service advertisement**: `_scrypted-viewport._tcp.local` on port 80. +- **SRV record**: hostname + port. +- **TXT records**: + - `version=1.0.0` + - `resolution=<effective>` (e.g. `480x800` for portrait, `800x480` for landscape) + - `orientation=<portrait|landscape>` + - `name=<viewport name>` (empty until `/config`) + +Scrypted-side discovery flow: + +1. Browse `_scrypted-viewport._tcp.local` via a Node mDNS-SD library (`bonjour-service`, `mdns`, etc.). These libraries talk to the multicast layer directly — they do **not** need OS-level `.local` resolution. +2. Each browse result contains `{name, host, port, addresses, txt}`. Use the **IP** from `addresses`, not the hostname, for all subsequent calls. +3. Match TXT `name=` against the operator's `viewport → camera` config bindings. +4. Re-browse periodically (every few minutes) to catch DHCP renumbering. + +If mDNS-SD is unavailable in the deployment (some Docker setups, certain VLAN configurations), allow the operator to set an explicit `http://<ip>:<port>` per viewport in Scrypted-side config as a fallback. ## API @@ -82,7 +96,7 @@ Returns `200 OK` with JSON: "last_frame_ms_ago": 1234, "frames_received": 4271, "decode_errors": 0, - "callback_failures": 2, + "state_post_failures": 2, "resolution": "480x800", "ip": "192.168.1.42", "free_heap": 123456, @@ -102,7 +116,7 @@ Returns `200 OK` with JSON: - Idempotent. Already-in-that-state calls return `204` and do nothing. - `wake`: backlight on, render loading screen (until the next `/frame`), reset idle timer. - `sleep`: backlight off; framebuffer discarded. -- No callback echo (Scrypted initiated). +- The device does NOT POST back to Scrypted (Scrypted initiated). - Response: `204`. ### GET /config @@ -111,22 +125,22 @@ Returns the persisted config: ```json { - "display": "mudroom", - "callback": "http://scrypted.local:11080/api/viewport/touch", + "viewport": "mudroom", + "scrypted": "http://scrypted.local:11080/endpoint/scrypted-viewport", "idle_timeout_ms": 60000, "orientation": "portrait", "brightness": 80 } ``` -Before first `/config`: returns `200` with an object whose fields are all `null` except defaults (`brightness: 80`, `orientation: "portrait"`, `idle_timeout_ms: 60000`). +Before first `/config`: returns `200` with `viewport` and `scrypted` as `null`; the rest carry their first-boot defaults (`brightness: 80`, `orientation: "portrait"`, `idle_timeout_ms: 60000`). ### POST /config ```json { - "display": "mudroom", - "callback": "http://scrypted.local:11080/api/viewport/touch", + "viewport": "mudroom", + "scrypted": "http://scrypted.local:11080/endpoint/scrypted-viewport", "idle_timeout_ms": 60000, "orientation": "portrait", "brightness": 80 @@ -136,7 +150,7 @@ Before first `/config`: returns `200` with an object whose fields are all `null` - **Partial update**: only fields present in the body are changed; omitted fields keep their current values. The persisted config is replaced atomically with the merged result. - Persisted to NVS, survives reboot. - Idempotent; reposting the same body yields the same state. -- `display` must be non-empty; `callback` must be `http://...`. +- `viewport` must be non-empty; `scrypted` must be `http://...` and is the Scrypted plugin's base URL. The device POSTs state changes to `<scrypted>/state`. - `idle_timeout_ms`: `0` disables the idle timer; non-zero values must be ≥ `5000`. Otherwise `400`. Scrypted should use the same value for its own per-stream timeout so both ends agree, but they time independently — either can end the session. - `orientation`: `portrait` (480x800) or `landscape` (800x480). Default `portrait` on first boot. Changing orientation takes effect immediately, including for the IP and Loading screens. Scrypted must send JPEGs at the new effective resolution after a change. - `brightness`: integer `0`–`100`. Default `80` on first boot. Applied immediately if awake; takes effect on next wake if asleep. PWM is gamma-corrected so the scale is perceptual. @@ -155,7 +169,7 @@ Paints a frame. Does **not** change wake/sleep state. - `Content-Type: image/jpeg`, body is raw JPEG bytes. - Image must match the effective resolution (480x800 portrait, 800x480 landscape) as a baseline JPEG. Device does not scale, rotate, or letterbox JPEG content. - Max size: 1 MB. -- **Requires awake state.** While asleep, returns `409 Conflict` and does not paint. Scrypted must `POST /state {"state":"wake"}` first (or wait for a `wake` callback from a tap). +- **Requires awake state.** While asleep, returns `409 Conflict` and does not paint. Scrypted must `POST /state {"state":"wake"}` first (or wait for a tap-driven `wake` POST from the device). - Resets the idle timer on success. - Single in-flight frame; concurrent posts may be rejected with `503`. - Returns `204` once decoded and pushed to the panel. @@ -163,48 +177,124 @@ Paints a frame. Does **not** change wake/sleep state. ## Wake / Sleep -Wake and sleep couple the device backlight with Scrypted's frame stream. The device owns the state; Scrypted just receives idempotent imperatives in callbacks: +Wake and sleep couple the device backlight with Scrypted's frame stream. The device owns the state. -- `state=wake` → "start streaming to this display now" -- `state=sleep` → "stop streaming to this display now" +Both the device and Scrypted expose the same endpoint, `POST /state`, with the same body shape `{viewport, state}` — they're peers. Either side can push to the other to set state. Repeats are safe; both sides are idempotent. -Scrypted does not track per-viewport state; it acts on the callback and forgets. Repeats are safe. +- `{"viewport": "<name>", "state": "wake"}` → "start streaming to this viewport now" +- `{"viewport": "<name>", "state": "sleep"}` → "stop streaming to this viewport now" + +Each request carries the device's `viewport` name as the routing key. Scrypted does not track per-viewport state across requests; it acts on each and forgets. Transitions: -| Trigger | Resulting state | Callback to Scrypted | +| Trigger | Resulting state | Device → Scrypted `POST /state`? | | --- | --- | --- | -| Tap while asleep | Awake (loading screen) | `state=wake event=tap` | -| Tap while awake | Asleep | `state=sleep event=tap` | -| Idle timer expires | Asleep | `state=sleep event=timeout` | +| Tap while asleep | Awake (loading screen) | `state=wake` | +| Tap while awake | Asleep | `state=sleep` | +| Idle timer expires | Asleep | `state=sleep` | | `POST /state {"state":"wake"}` | Awake (loading screen) | none | | `POST /state {"state":"sleep"}` | Asleep | none | | `POST /frame` | (no state change) | — | -`/frame` never changes state. Scrypted must `POST /state {"state":"wake"}` (or wait for a tap-driven callback) before sending frames. This makes the protocol race-free: a `/frame` arriving after a tap-to-sleep is rejected with `409`, not silently re-woken. +`/frame` never changes state. Scrypted must `POST /state {"state":"wake"}` (or wait for a tap-driven `wake` POST from the device) before sending frames. This makes the protocol race-free: a `/frame` arriving after a tap-to-sleep is rejected with `409`, not silently re-woken. Only `tap` is detected on the touchscreen — long-press and swipes are out of scope for v1. `tap` itself is internal; what Scrypted sees is the resulting `state`. -Callback body: +### Device → Scrypted `POST /state` + +When the device changes state on its own (a tap, the idle timer firing), it POSTs to Scrypted's `/state` endpoint. Same shape as Scrypted POSTing to the device's `/state`; no `event`, no `type`, no callback semantics. The two endpoints are peers, not request/response. + +**Request** + +``` +POST <scrypted>/state HTTP/1.1 +Host: <derived from URL> +Content-Type: application/json +Content-Length: <body length> +User-Agent: ScryptedViewport/<version> +Connection: close +``` + +Body: ```json -{ - "display": "mudroom", - "state": "wake", - "event": "tap" -} +{ "viewport": "mudroom", "state": "wake" } +``` + +or + +```json +{ "viewport": "mudroom", "state": "sleep" } ``` -- `state`: `wake` or `sleep` — the resulting state, also the imperative for Scrypted. -- `event`: the cause. `tap` (any user tap) or `timeout` (idle expiry, sleep only). Future events (e.g. `swipe_left`) can be added without schema changes. Scrypted may ignore `event` in v1. +Same body as Scrypted → Device `POST /state`, plus the routing key. + +- `viewport` (string): the value of the `viewport` field in the device's `/config`. Scrypted's routing key. +- `state` (string): `wake` or `sleep`. The resulting state and the imperative for Scrypted. + +No other fields. No timestamp (no RTC, no SNTP); Scrypted timestamps on receipt. + +**No application-level ack** + +HTTP 2xx is transport-level only. There is no application-level ack: the device does not retry, does not block subsequent state changes on the response, and does not treat a 5xx response as anything more than a counter increment. Idempotency + `/frame` returning `409` + each side's independent idle timer recover every failure mode without an ack: + +- Scrypted misses the device's `sleep`: its own per-stream timer eventually stops the stream, or the next `/frame` it sends returns `409`. +- Device misses Scrypted's `wake`/`sleep`: same — the next state change on either side syncs them. -Delivery: best-effort, ~1s timeout, no retry. Callbacks before `/config` registers a URL are dropped. If the callback POST fails, the local state change still happens — Scrypted catches up via its own timeout or the next callback. +Don't design Scrypted-side logic that waits for the device to confirm a state change. There is no confirmation. -No wall-clock timestamp is included. The device has no RTC and no SNTP. Scrypted timestamps callbacks on receipt. +**Expected response (counter only)** + +- Any 2xx is success. +- Body is ignored. +- Anything else (non-2xx, connection refused, DNS failure, request timeout) increments `state_post_failures` and is otherwise ignored. + +**Timeouts** + +- Connect timeout: 1s. +- Total request timeout: 1s. +- After timeout the device aborts the connection and moves on. + +**Concurrency and ordering** + +- At most one outbound `/state` POST in flight at a time. +- POSTs are delivered in the order state changes occur on the device. +- If a state change happens while a POST is in flight, it goes into a depth-1 queue. If the queue already holds a POST, the queued entry is **replaced** by the newer one. The in-flight POST is never cancelled. +- Replacement is safe because POSTs are imperatives: only the latest desired state matters to Scrypted. Intermediate flips between wake and sleep within the queue window collapse to the final state. + +**Failure semantics** + +- The local state change always happens regardless of POST outcome. +- A dropped or failed POST is recovered by the next user tap, the device's idle timer firing `sleep`, or the next `/frame` from Scrypted returning `409`. +- No retry queue. No backoff. No persistence across reboots. + +**When the device does NOT POST** + +- Before `/config` has registered a `scrypted` URL (boot state, factory reset). Silently dropped. +- For state changes Scrypted initiated (`POST /state`, `POST /frame` while asleep). Scrypted already knows; echoing would loop. +- BOOT short-press IP overlay (not a state change). + +**Trust model** + +Plain HTTP, no TLS, no auth — same as the inbound API. LAN-only. + +### Race handling + +Wake/sleep changes can race: a user taps the device while Scrypted is mid-flight with a `POST /state` from a stale camera-event timeout, or the idle timer fires at the same instant Scrypted POSTs a fresh `wake`. The protocol does **not** carry epochs, session IDs, or priorities. Race resolution is purely about each side serializing its own writes and trusting idempotency to converge. + +Rules: + +1. **Device-side serialization.** The device guards its state-mutation function with a mutex. Tap, idle timer, and `POST /state` all funnel through it. Whichever lands second wins. +2. **Scrypted-side serialization.** Scrypted handles inbound `POST /state` requests for a given `viewport` one at a time (in-process queue per viewport). Whichever lands second wins. +3. **Last write wins.** No priorities. No "the device is where the user is so it always trumps Scrypted." If a stale `sleep` lands after a fresh `wake`, the device sleeps; the user taps again and we're back. One extra tap is cheap. +4. **Scrypted must cancel its own pending operations on each inbound POST.** When a `wake` arrives, Scrypted cancels any pending per-viewport sleep timer before starting a fresh stream. Same in reverse. This makes "stale Scrypted timer fires after the user tapped to wake" impossible without needing protocol-level epochs. + +Idempotency does the rest — `POST /state` on either side and `POST /frame` (relative to its `409`-vs-`204` behavior) are all no-ops when the recipient is already in the requested state. ## Idle -After `idle_timeout_ms` (default 60s) with no `/frame`, the device sleeps and POSTs `state=sleep` with `event=timeout`. Scrypted should use the same timeout so its per-stream cutoff matches, but they run independently — either side can end the session, whichever notices first. +After `idle_timeout_ms` (default 60s) with no `/frame`, the device sleeps and POSTs `state=sleep`. Scrypted should use the same timeout so its per-stream cutoff matches, but they run independently — either side can end the session, whichever notices first. ## Idempotency @@ -218,18 +308,18 @@ All endpoints are safe to retry. Every state-change path converges to the same f | `POST /state` | yes | no-op if already in that state | | `POST /frame` | yes (within state) | paints if awake; `409` if asleep — no partial state | -Callbacks (`state=wake`, `state=sleep`) are imperatives, not notifications. Scrypted acts and forgets; the device never expects an ack. A dropped callback is recovered by the next user action, by Scrypted's own timeout, or by a `/frame` returning `409`. +`POST /state` (in either direction) carries imperatives, not notifications. Both sides act and forget; neither expects an application-level ack. A dropped POST is recovered by the next user action, by Scrypted's own timeout, or by a `/frame` returning `409`. Failure modes do not corrupt state: -- Failed callback POST: local state still changes; counter increments. +- Failed outbound `/state` POST: local state still changes; `state_post_failures` increments. - Failed JPEG decode: previous frame stays; state unchanged. - Network drop mid-stream: device idle-sleeps when the timer expires. - Concurrent `/frame` posts: one wins, the other gets `503`; no half-painted frames. ## BOOT button -- **Short press**: overlay the IP screen for 15 seconds, then return to the prior state. Useful for identifying or re-registering a device that's already configured. Wakes the backlight temporarily; does not change the wake/sleep state or send a callback. An incoming `/frame` while the overlay is showing is rejected with `409` (state is still "sleep" underneath). +- **Short press**: overlay the IP screen for 15 seconds, then return to the prior state. Useful for identifying or re-registering a device that's already configured. Wakes the backlight temporarily; does not change the wake/sleep state and does not POST to Scrypted. An incoming `/frame` while the overlay is showing is rejected with `409` (state is still "sleep" underneath). - **Hold ≥5s**: factory reset — clear NVS, reboot. The device comes back unconfigured, showing the IP screen until Scrypted POSTs `/config`. ## Local rendering @@ -245,7 +335,7 @@ Both use a small embedded bitmap font. No LVGL, no general text engine. The Scrypted side is **code, not configuration** — Scrypted has no built-in concept of a network framebuffer. The code is small and lives inside Scrypted: -- **v1**: a Scrypted Script (in the Scripts plugin) — listens for camera events, calls `takePicture()`, POSTs the JPEG to `/frame`. Exposes the `/api/viewport/touch` endpoint via the EndpointManager. ~50 lines of TypeScript, no package install. +- **v1**: a Scrypted Script (in the Scripts plugin) — listens for camera events, calls `takePicture()`, POSTs the JPEG to `/frame`. Exposes a `POST /state` handler at the plugin's endpoint root (e.g. `http://scrypted.local:11080/endpoint/scrypted-viewport/state`) via the EndpointManager. ~50 lines of TypeScript, no package install. - **Next milestone after v1 (`/stream`)**: add `POST /stream` with `multipart/x-mixed-replace` chunked body for live frame rates. Scrypted side becomes a small custom plugin using FFmpeg via `MediaManager` to pipe MJPEG. `/frame` stays for snapshots and debug. Either way, no Scrypted core changes and no external service. @@ -255,12 +345,14 @@ Scrypted owns a static list of viewports, each **bound to one Scrypted camera de On startup, register every viewport: ```ts -await fetch(`${viewport}/config`, { +const SCRYPTED_BASE = "http://scrypted.local:11080/endpoint/scrypted-viewport"; + +await fetch(`${v.url}/config`, { method: "POST", headers: { "Content-Type": "application/json" }, body: JSON.stringify({ - display: "mudroom", - callback: "http://scrypted.local:11080/api/viewport/touch", + viewport: v.name, // e.g. "mudroom" + scrypted: SCRYPTED_BASE, idle_timeout_ms: 60000, orientation: "portrait", brightness: 80 @@ -271,35 +363,48 @@ await fetch(`${viewport}/config`, { To start a session (camera event like doorbell, motion, person): ```ts -await fetch(`${viewport}/state`, { +await fetch(`${v.url}/state`, { method: "POST", headers: { "Content-Type": "application/json" }, body: JSON.stringify({ state: "wake" }) }); // device shows loading // then stream frames: -await fetch(`${viewport}/frame`, { +await fetch(`${v.url}/frame`, { method: "POST", headers: { "Content-Type": "image/jpeg" }, body: jpegBuffer // baseline JPEG at the viewport's effective resolution, <1 MB }); ``` -Handle callbacks at `POST /api/viewport/touch` (body: `{display, state, event}`): +Expose `POST /state` at `<SCRYPTED_BASE>/state` (peer of the device's `/state`). Body is `{viewport, state}`: + +```ts +const { viewport, state } = req.body; +const v = viewports.find(x => x.name === viewport); +if (!v) return res.status(404).end(); + +cancelPendingSleep(v); // race rule: cancel stale timers on every incoming POST + +if (state === "wake") startStream(v); // idempotent +if (state === "sleep") stopStream(v); // idempotent + +res.status(204).end(); +``` -- `state=wake` → look up the camera bound to `display`, start streaming frames. (You do not need to `POST /state` first — the device is already awake when it sends this.) -- `state=sleep` → stop streaming frames to `display`. +- `state=wake` → start streaming frames to the viewport's bound camera. You do not need to `POST /state` to the device first; the device is already awake when it sends this. +- `state=sleep` → stop streaming frames to that viewport. -Both are idempotent on Scrypted's side. Don't track viewport state; act on the callback and forget. Ignore the `event` field in v1. +Both are idempotent. Don't track viewport state across requests; act on each and forget. -Scrypted should use the same `idle_timeout_ms` value it sent in `/config` as its own per-stream cutoff. The two timers run independently — either side can cut a session, whichever notices first. If the viewport's sleep callback is lost, the Scrypted-side timeout still ends the stream; the next `/frame` posted after the device idle-slept simply returns `409` and Scrypted stops. +Scrypted should use the same `idle_timeout_ms` value it sent in `/config` as its own per-stream cutoff. The two timers run independently — either side can cut a session, whichever notices first. If the device's outbound `sleep` POST is lost, the Scrypted-side timeout still ends the stream; the next `/frame` posted after the device idle-slept simply returns `409` and Scrypted stops. ## Ops - Firmware updates: reflash over USB. No OTA in v1 (planned post-v1: HTTP OTA from Scrypted). - Provisioning: flash the same firmware to every device. On first boot the screen shows its IP; register it from Scrypted via `POST /config`. -- Display names must be unique across the LAN — mDNS hostnames are derived from `display` and two viewports configured with the same name will collide. -- Factory reset: hold BOOT for 5s during normal operation to clear NVS (display name, callback, brightness, idle timeout, orientation) and reboot. The device returns to the IP screen. +- Viewport names must be unique across the LAN — mDNS hostnames are derived from `viewport` and two devices configured with the same name will collide. +- Factory reset: hold BOOT for 5s during normal operation to clear NVS (viewport name, scrypted URL, brightness, idle timeout, orientation) and reboot. The device returns to the IP screen. - No DHCP lease: keep retrying; do not reboot. Screen shows "no network" if unconfigured. - Ethernet disconnect: reconnect automatically. If Scrypted is unreachable, displays go stale — nothing the device can do about it. - Watchdog: the ESP-IDF task watchdog reboots the device if a task hangs. Soft state is rebuilt from NVS on every boot. @@ -322,10 +427,10 @@ Scrypted Viewport is a thin network framebuffer appliance. ESP: - DHCP - mDNS -- HTTP server +- HTTP server (`/state`, `/config`, `/frame`) +- HTTP client (outbound `/state` POSTs to Scrypted) - JPEG decode - framebuffer - touch -- callback Everything else belongs in Scrypted. |
