feat: add support for passing latency to generate_stream_input si…

…milar to `generate_stream` (elevenlabs#116) Co-authored-by: Flavio Schneider <flavio.schneider.97@gmail.com>
Bit-o-lodeon-Ai · Dec 8, 2023 · 7d25b2e · 7d25b2e
1 parent d08ef67
commit 7d25b2e
Show file tree

Hide file tree

Showing 2 changed files with 3 additions and 2 deletions.
diff --git a/elevenlabs/api/tts.py b/elevenlabs/api/tts.py
@@ -89,6 +89,7 @@ def generate_stream_input(
         model: Model,
         api_key: Optional[str] = None,
         output_format: TTS.OutputFormat = "mp3_44100_128",
+        latency: int = 1,
     ) -> Iterator[bytes]:
         BOS = json.dumps(
             dict(
@@ -103,7 +104,7 @@ def generate_stream_input(
         EOS = json.dumps(dict(text=""))
 
         with connect(
-            f"wss://api.elevenlabs.io/v1/text-to-speech/{voice.voice_id}/stream-input?model_id={model.model_id}&output_format={output_format}",
+            f"wss://api.elevenlabs.io/v1/text-to-speech/{voice.voice_id}/stream-input?model_id={model.model_id}&output_format={output_format}&optimize_streaming_latency={latency}",
             additional_headers={
                 "xi-api-key": api_key or os.environ.get("ELEVEN_API_KEY")
             },

diff --git a/elevenlabs/simple.py b/elevenlabs/simple.py
@@ -68,7 +68,7 @@ def generate(
                 text, voice, model, stream_chunk_size, api_key=api_key, latency=latency, output_format=output_format
             )  # noqa E501
         elif isinstance(text, Iterator):
-            return TTS.generate_stream_input(text, voice, model, api_key=api_key, output_format=output_format)
+            return TTS.generate_stream_input(text, voice, model, api_key=api_key, output_format=output_format, latency=latency)
     else:
         assert isinstance(text, str)
         return TTS.generate(text, voice, model, api_key=api_key, output_format=output_format)