Merge pull request #1671 from pipecat-ai/khk/rime-arcana

support for rime arcana model

CHANGELOG.md

@@ -5,6 +5,12 @@ All notable changes to **Pipecat** will be documented in this file.
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
+## [Unreleased]
+
+### Added
+
+- Added support to `RimeHttpTTSService` for the `arcana` model.
+
 ## [0.0.66] - 2025-05-02
 
 ### Added
@@ -120,7 +126,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
   not a very common case.
 
 - Added `RivaSegmentedSTTService`, which allows Riva offline/batch models, such
-  as  to be "canary-1b-asr" used in Pipecat.
+  as to be "canary-1b-asr" used in Pipecat.
 
 ### Deprecated
 

examples/foundational/07q-interruptible-rime-http.py

@@ -44,7 +44,8 @@ async def run_bot(webrtc_connection: SmallWebRTCConnection, _: argparse.Namespac
 
         tts = RimeHttpTTSService(
             api_key=os.getenv("RIME_API_KEY", ""),
-            voice_id="rex",
+            voice_id="luna",
+            model="arcana",
             aiohttp_session=session,
         )
 

src/pipecat/services/rime/tts.py

@@ -400,6 +400,13 @@ class RimeHttpTTSService(TTSService):
         payload["modelId"] = self._model_name
         payload["samplingRate"] = self.sample_rate
 
+        # Arcana does not support PCM audio
+        if payload["modelId"] == "arcana":
+            headers["Accept"] = "audio/wav"
+            need_to_strip_wav_header = True
+        else:
+            need_to_strip_wav_header = False
+
         try:
             await self.start_ttfb_metrics()
 
@@ -420,6 +427,10 @@ class RimeHttpTTSService(TTSService):
                 CHUNK_SIZE = 1024
 
                 async for chunk in response.content.iter_chunked(CHUNK_SIZE):
+                    if need_to_strip_wav_header and chunk.startswith(b"RIFF"):
+                        chunk = chunk[44:]
+                        need_to_strip_wav_header = False
+
                     if len(chunk) > 0:
                         await self.stop_ttfb_metrics()
                         frame = TTSAudioRawFrame(chunk, self.sample_rate, 1)