@origonai/web-sdk 0.1.2 → 0.1.3

package/docs/backend-integration.md

@@ -82,7 +82,7 @@ Document a min-supported-browser policy with the product team.
 |---|---|---|
 | Bidi QUIC stream | `wt.createBidirectionalStream()` (client-opened) | ControlMessages — Connect, NegotiateMedia, Hold, Mute, SendDtmf, Resume + their `*Ok` + server-push |
 | QUIC datagrams | `wt.datagrams` (readable + writable streams) | Audio frames (one frame per datagram), uplink + downlink |
-| QUIC uni stream (server→client) | `wt.incomingUnidirectionalStreams` | DTMF events when the peer pressed a digit |
+| QUIC uni stream (server→client) | `wt.incomingUnidirectionalStreams` | Drained as unsupported protocol input; no public receive lane |
 | QUIC graceful close | `wt.close({ closeCode, reason })` | Hangup |
 
 Everything in [`client-integration.md`](client-integration.md) §4–§12 applies
@@ -154,8 +154,8 @@ After `ConnectOk` the bidi multiplexes:
 - Your verbs out (`Hold`, `Mute`, `SendDtmf`, `Resume`, `NegotiateMedia`).
 - Server `*Ok` responses (correlate by `request_id`).
 - Server-push events (`PeerAttached`, `PeerDetached`, `EndpointHeld`,
-  `EndpointMuted`, `PeerStateChanged`, `DtmfReceived`, `EndpointHangup`,
-  `MoqError`) — these carry `request_id = 0`.
+  `EndpointMuted`, `PeerStateChanged`, `EndpointHangup`, `MoqError`) — these
+  carry `request_id = 0`.
 
 A single read loop on `ctrl.readable.getReader()` is sufficient. Dispatch
 by `msg_type`.
@@ -232,30 +232,18 @@ manual PLC). Server never retransmits.
 
 ---
 
-## 6. DTMF inbound (unidirectional sub-stream)
+## 6. DTMF flow input
 
-DTMF *from a peer* arrives on a fresh server-initiated uni stream, one
-per digit:
+`SessionManager.sendDtmf(sessionId, digit)` sends exactly one symbol from
+`0-9*#A-D` as `SendDtmf` on the bidi control stream. It resolves only after
+the matching `SendDtmfOk`, which acknowledges Media publication to the active
+CX flow. Request-scoped Media errors reject only the matching call.
 
-```js
-const uniReader = wt.incomingUnidirectionalStreams.getReader()
-while (true) {
-  const { value: stream, done } = await uniReader.read()
-  if (done) break
-  // stream is a ReadableStream<Uint8Array>
-  handleDtmfUniStream(stream)        // decodes SubgroupHeader + DtmfEvent + EndOfGroup
-}
-```
-
-The framing on this stream is documented in
-[`docs/wire-protocol.md`](wire-protocol.md) §"Stream paths (non-audio)".
-The payload is a single `DtmfEvent { digit, duration_ms }` protobuf
-message.
-
-DTMF *to the server* (you pressing a digit) is the `SendDtmf` verb on
-the bidi control stream — not a uni stream. The browser cannot open
-WT uni sub-streams at all in some specs/browsers, so we never required
-it for the outbound direction.
+The request belongs to the WebTransport on which it was sent. A close rejects
+it; reconnect/Resume never queues or replays it, and a stale ACK is ignored.
+The SDK does not receive peer-generated DTMF, expose a DTMF callback, synthesize
+a tone, or imply PSTN/carrier delivery. Incoming uni streams are drained only
+to keep the transport healthy.
 
 ---
 
@@ -268,13 +256,12 @@ Same shape as native. See [`client-integration.md`](client-integration.md)
 |---|---|---|
 | `Hold { held: bool }` | Client → Server | `session.hold` |
 | `Mute { scope: Uplink \| Downlink \| Both \| None }` | Client → Server | `session.mute_self` |
-| `SendDtmf { digit, duration_ms }` | Client → Server | `dtmf_send` |
+| `SendDtmf { request_id, digit }` | Client → Server | `dtmf_send` |
 | `Resume { call_token, last_event_seq }` | Client → Server | (token only) |
 | `NegotiateMedia { offers, frame_size_ms }` | Client → Server | (token only; additive) |
 
-The corresponding `*Ok` responses carry the echoed `request_id`. Server
-pushes the matching state-change event (`EndpointHeld`, `EndpointMuted`,
-`PeerStateChanged`, `DtmfReceived`) for other peers' UI consistency.
+The corresponding `*Ok` responses carry the echoed `request_id`. DTMF has no
+peer receive event; `SendDtmfOk` acknowledges Media publication only.
 
 ---
 
@@ -446,7 +433,7 @@ async function startCall({ url, jwt, certHash /* optional */ }) {
   startBidiReadLoop(reader, dispatchEvent)              // server-push + verb Oks
   startDatagramSender(wt, audioAlias, serverEpochUnix)  // mic → Opus → datagram
   startDatagramReceiver(wt)                              // datagrams → decoders
-  startUniReceiver(wt)                                   // DTMF uni streams
+  startUniStreamDrain(wt)                                // unsupported server uni input
 
   // 5. Wait for session end
   wt.closed.then(info => onClosed(info), err => onError(err))

package/docs/contract.md

@@ -39,6 +39,27 @@ Run typecheck, unit tests, and the production build, then rebuild both linked
 declarations and artifact bytes before publication. Publishing to npm remains
 an owner action and is never part of an implementation spin.
 
+## Internal voice DTMF wire codec
+
+- The browser codec encodes `SEND_DTMF=0x58` as
+  `requestId (QUIC varint) | digit (exactly one ASCII byte)` and decodes
+  `SEND_DTMF_OK=0x59` as the echoed request id only. The accepted alphabet is
+  exactly `0-9*#A-D`; empty, multi-byte, lowercase, non-ASCII, and trailing
+  bytes fail closed.
+- Request-scoped Media refusals remain the existing `MoqError` path:
+  `CapabilityMissing`, `IllegalState`, or `ResourceExhausted`. An ACK means
+  input acceptance for the owning CX flow, not peer delivery, tone playback,
+  or carrier forwarding.
+- `src/voice/wire/testdata/send_dtmf_golden.txt` is a byte-identical copy of
+  Media's producer-owned canonical fixture at
+  `core/media/modules/moq-endpoint/testdata/send_dtmf_golden.txt`.
+- `VoiceSession.sendDtmf(digit)` and
+  `SessionManager.sendDtmf(sessionId, digit)` expose the send-only API. Calls
+  reject during Resume, pending requests reject on transport close, and no
+  request is replayed on a replacement transport. There is no peer-receive
+  callback or tone playback. Media error-code changes and canonical fixture
+  changes therefore require a web-sdk consumer review.
+
 ## Live-chat monitoring and optional message metadata
 
 - Version 0.1.2 exposes optional `Message.metadata` and optional

package/docs/voice.md

@@ -5,7 +5,7 @@
 | Layer        | Tech                                                                 |
 |--------------|----------------------------------------------------------------------|
 | Transport    | WebTransport (`https://<media-server>/web`), one bidi control sub-stream + datagrams |
-| Control      | B2BUA verbs (Connect/ConnectOk/Mute/Resume/PeerAttached/…), `msg_type (vi) \| payload_len (vi) \| payload` |
+| Control      | B2BUA verbs (Connect/ConnectOk/Mute/Resume/SendDtmf/PeerAttached/…), `msg_type (vi) \| payload_len (vi) \| payload` |
 | Audio        | Opus 48 kHz mono, 20 ms frames (50 per wall-second), `ObjectDatagram` per frame |
 | Group / object IDs | `group_id = floor(Date.now()/1000) - server_epoch_unix`, `object_id ∈ [0..49]`, `end_of_group` on object 49 |
 | Reconnect    | Tier-3 `Resume { call_token, last_event_seq }` only — browsers expose no QUIC migration / 0-RTT |
@@ -13,6 +13,10 @@
 
 See [`backend-integration.md`](backend-integration.md) (canonical spec) for the full wire definitions.
 
+The internal `SendDtmf` codec carries one `0-9*#A-D` symbol and awaits its
+request-correlated ACK. It does not yet expose a public send method, receive
+callback, tone generator, duration control, or carrier-forwarding behavior.
+
 ## Required browser features (the SDK-wide floor)
 
 | Feature                | Needed by                                        |
@@ -128,6 +132,7 @@ const { sessionId } = await client.joinSession({
 })
 
 await client.setMute(sessionId, 'uplink')   // 'uplink' | 'downlink' | 'both' | 'none'
+await client.sendDtmf(sessionId, '#')       // exactly one of 0-9, *, #, A-D
 
 // For a provisioned receive-only join, enter and leave microphone capture
 // without exposing an unmuted permission/acquisition window.
@@ -147,6 +152,12 @@ await client.endSession(sessionId)
 
 ## Reconnect behaviour
 
+DTMF requests are transport-bound. Calls made during reconnect reject locally;
+an unacknowledged digit is rejected when its transport closes and is never
+replayed on the replacement transport. `SendDtmfOk` means CX-flow publication
+was accepted, not carrier delivery. There is no inbound DTMF callback or SDK
+tone playback.
+
 When `wt.closed` fires unexpectedly (no `EndpointHangup` or terminal `MoqError` arrived first), the SDK runs:
 
 ```

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@origonai/web-sdk",
-  "version": "0.1.2",
+  "version": "0.1.3",
   "description": "Origon Web SDK - chat and voice session client",
   "type": "module",
   "packageManager": "pnpm@11.5.1",