feat(rime): support Coda streaming in LiveKit

livekit-plugins/livekit-plugins-rime/README.md

@@ -13,3 +13,53 @@ pip install livekit-plugins-rime
 ## Pre-requisites
 
 You'll need an API key from Rime. It can be set as an environment variable: `RIME_API_KEY`
+
+## Streaming Coda WebSocket API
+
+The Rime v1 WebSocket protocol accepts streaming text and returns audio before the input turn
+is complete. The plugin aggregates input fragments into complete sentences before it sends them
+to Coda. All sentences in one LiveKit output turn use one continuous Coda context.
+
+```python
+import os
+
+from livekit.plugins import rime
+
+tts = rime.TTS(
+    websocket_url="wss://api.rime.ai/coda/ws",
+    speaker="astra",
+    api_key=os.environ["RIME_API_KEY"],
+)
+```
+
+Pass the active Coda WebSocket endpoint explicitly. The presence of `websocket_url` selects Coda,
+WebSocket streaming, and the v1 JSON protocol. The speaker defaults to `astra`. The plugin uses
+`livekit.agents.tokenize.blingfire.SentenceTokenizer` by default and configures it to emit one
+complete sentence at a time. Pass `tokenizer` to select another LiveKit sentence tokenizer. A
+custom tokenizer must emit complete sentence units that are safe for Coda text normalization.
+
+One LiveKit stream uses one continuous Rime synthesis context. These stream methods map to the
+Rime lifecycle as follows:
+
+| LiveKit method | Local action | Rime operation |
+| --- | --- | --- |
+| `stream.push_text()` | Buffer and sentence-tokenize text. | Send `text` for each completed sentence. |
+| `stream.flush()` | Release the current tokenizer buffer and keep the context open. | Send released content as `text`; no `flush` operation exists. |
+| `stream.end_input()` | Drain final text and finalize input. | Send `end`. |
+| `stream.aclose()` | Stop active synthesis. | Send `cancel` when needed. |
+
+You can send more text after `flush()`. The same Coda context remains active, and an input pause
+needs no wire message. Only `end_input()` ends normal input and causes Coda to send `done`.
+Each outgoing `text` value must be a stable sentence-sized unit. Calling `flush()` drains any
+buffered fragment, so call it only after a complete sentence or a stable clause.
+
+The first v1 implementation has these limits:
+
+- WebSocket v1 supports the `coda` model only.
+- It requests raw `audio/pcm` data.
+- It does not provide aligned word timestamps.
+- `speed_alpha` is not supported. Use `time_scale_factor` to control speed.
+- The adapter uses the JSON `rime.v1.json` WebSocket subprotocol.
+- The adapter reuses WebSocket connections between sequential LiveKit streams. It does not run
+  concurrent Coda contexts on one WebSocket. Concurrent LiveKit streams use separate pooled
+  WebSocket connections.