[email protected]

Patch Changes

• use quantized onnx version of turn detector model - #1231 (@jeradf)

• use onnxruntime for turn detection and remove pytorch dependency - #1257 (@jeradf)

[email protected]

Patch Changes

• Moved create_ai_function_info to function_context.py for better reusability and reduce repetation - #1260 (@jayeshp19)

• add on_duplicate option for multimodal agent response create - #1204 (@longcw)

• Add support for OpenAI's "detail" parameter to ChatImage - #1213 (@bcherry)

  Add support for data URLs on ChatImage in the Anthropic plugin.

• filter out empty message for set chat ctx in realtime model - #1245 (@longcw)

• fix: correctly parse function argument types - #1221 (@jayeshp19)

• add session_updated event for RealtimeSession - #1253 (@longcw)

• added llama 3.3 70b to model definitions - #1233 (@davidzhao)

• update default realtime model to gpt-4o-realtime-preview-2024-12-17 - #1250 (@davidzhao)

• Fix center_aspect_fit bug, add scale_aspect_fit and scale_aspect_fill resizing options. - #1222 (@bcherry)

  Make scale_aspect_fit the new default resizing option for video frames.

[email protected]

Patch Changes

• added streaming audio decoder for compressed audio. - #1236 (@davidzhao)

• Support Deepgram TTS - #1201 (@jayeshp19)

[email protected]

Patch Changes

• fix: fix imgui setup - #1226 (@mbukeRepo)

[email protected]

Patch Changes

• fix azure stt language autodetection - #1246 (@davidzhao)

[email protected]

Patch Changes

• Moved create_ai_function_info to function_context.py for better reusability and reduce repetation - #1260 (@jayeshp19)

• Add support for OpenAI's "detail" parameter to ChatImage - #1213 (@bcherry)

  Add support for data URLs on ChatImage in the Anthropic plugin.

• fix: correctly parse function argument types - #1221 (@jayeshp19)

• Fix center_aspect_fit bug, add scale_aspect_fit and scale_aspect_fill resizing options. - #1222 (@bcherry)

  Make scale_aspect_fit the new default resizing option for video frames.

[email protected]

Patch Changes

• Moved create_ai_function_info to function_context.py for better reusability and reduce repetation - #1260 (@jayeshp19)

• added streaming audio decoder for compressed audio. - #1236 (@davidzhao)

• Add JPEG quality param to image encoder - #1249 (@bcherry)

• Add support for OpenAI's "detail" parameter to ChatImage - #1213 (@bcherry)

  Add support for data URLs on ChatImage in the Anthropic plugin.

• fix: correctly parse function argument types - #1221 (@jayeshp19)

• Fix center_aspect_fit bug, add scale_aspect_fit and scale_aspect_fill resizing options. - #1222 (@bcherry)

  Make scale_aspect_fit the new default resizing option for video frames.

[email protected]

Patch Changes

• improvements to endpointing latency - #1212 (@davidzhao)

• Improvements to end of turn plugin, ensure STT language settings. - #1195 (@davidzhao)

[email protected]

Patch Changes

• add google/gemini-2.0-flash-exp as default model for vertex - #1214 (@jayeshp19)

• emit error event for realtime model - #1200 (@longcw)

• fix: return structured output from func calls - #1187 (@jayeshp19)

• Handle optional func args in tool calls when set to None - #1211 (@jayeshp19)

• fix: openai llm retries - #1196 (@theomonnom)

• Improvements to end of turn plugin, ensure STT language settings. - #1195 (@davidzhao)

• fix: Handle optional func args in tool calls when set to None - #1211 (@jayeshp19)

[email protected]

Patch Changes

• enable deepgram filler words by default to improve end of turn accuracy - #1190 (@davidzhao)