ServiceAttributes

Functions for adding attributes to OpenTelemetry spans.

pipecat.utils.tracing.service_attributes.add_tts_span_attributes(span, service_name, model, voice_id, text=None, settings=None, character_count=None, operation_name='tts', ttfb=None, **kwargs)[source]

Add TTS-specific attributes to a span.

Parameters:

  • span (Span) – The span to add attributes to

  • service_name (str) – Name of the TTS service (e.g., “cartesia”)

  • model (str) – Model name/identifier

  • voice_id (str) – Voice identifier

  • text (str | None) – The text being synthesized

  • settings (Dict[str, Any] | None) – Service configuration settings

  • character_count (int | None) – Number of characters in the text

  • operation_name (str) – Name of the operation (default: “tts”)

  • ttfb (float | None) – Time to first byte in seconds

  • **kwargs – Additional attributes to add

Return type:

None

pipecat.utils.tracing.service_attributes.add_stt_span_attributes(span, service_name, model, operation_name='stt', transcript=None, is_final=None, language=None, settings=None, vad_enabled=False, ttfb=None, **kwargs)[source]

Add STT-specific attributes to a span.

Parameters:

  • span (Span) – The span to add attributes to

  • service_name (str) – Name of the STT service (e.g., “deepgram”)

  • model (str) – Model name/identifier

  • operation_name (str) – Name of the operation (default: “stt”)

  • transcript (str | None) – The transcribed text

  • is_final (bool | None) – Whether this is a final transcript

  • language (str | None) – Detected or configured language

  • settings (Dict[str, Any] | None) – Service configuration settings

  • vad_enabled (bool) – Whether voice activity detection is enabled

  • ttfb (float | None) – Time to first byte in seconds

  • **kwargs – Additional attributes to add

Return type:

None

pipecat.utils.tracing.service_attributes.add_llm_span_attributes(span, service_name, model, stream=True, messages=None, output=None, tools=None, tool_count=None, tool_choice=None, system=None, parameters=None, extra_parameters=None, ttfb=None, **kwargs)[source]

Add LLM-specific attributes to a span.

Parameters:

  • span (Span) – The span to add attributes to

  • service_name (str) – Name of the LLM service (e.g., “openai”)

  • model (str) – Model name/identifier

  • stream (bool) – Whether streaming is enabled

  • messages (str | None) – JSON-serialized messages

  • output (str | None) – Aggregated output text from the LLM

  • tools (str | None) – JSON-serialized tools configuration

  • tool_count (int | None) – Number of tools available

  • tool_choice (str | None) – Tool selection configuration

  • system (str | None) – System message

  • parameters (Dict[str, Any] | None) – Service parameters

  • extra_parameters (Dict[str, Any] | None) – Additional parameters

  • ttfb (float | None) – Time to first byte in seconds

  • **kwargs – Additional attributes to add

Return type:

None

pipecat.utils.tracing.service_attributes.add_gemini_live_span_attributes(span, service_name, model, operation_name, voice_id=None, language=None, modalities=None, settings=None, tools=None, tools_serialized=None, transcript=None, is_input=None, text_output=None, audio_data_size=None, **kwargs)[source]

Add Gemini Live specific attributes to a span.

Parameters:

  • span (Span) – The span to add attributes to

  • service_name (str) – Name of the service

  • model (str) – Model name/identifier

  • operation_name (str) – Name of the operation (setup, model_turn, tool_call, etc.)

  • voice_id (str | None) – Voice identifier used for output

  • language (str | None) – Language code for the session

  • modalities (str | None) – Supported modalities (e.g., “AUDIO”, “TEXT”)

  • settings (Dict[str, Any] | None) – Service configuration settings

  • tools (List[Dict] | None) – Available tools/functions list

  • tools_serialized (str | None) – JSON-serialized tools for detailed inspection

  • transcript (str | None) – Transcription text

  • is_input (bool | None) – Whether transcript is input (True) or output (False)

  • text_output (str | None) – Text output from model

  • audio_data_size (int | None) – Size of audio data in bytes

  • **kwargs – Additional attributes to add

Return type:

None

pipecat.utils.tracing.service_attributes.add_openai_realtime_span_attributes(span, service_name, model, operation_name, session_properties=None, transcript=None, is_input=None, context_messages=None, function_calls=None, tools=None, tools_serialized=None, audio_data_size=None, **kwargs)[source]

Add OpenAI Realtime specific attributes to a span.

Parameters:

  • span (Span) – The span to add attributes to

  • service_name (str) – Name of the service

  • model (str) – Model name/identifier

  • operation_name (str) – Name of the operation (setup, transcription, response, etc.)

  • session_properties (Dict[str, Any] | None) – Session configuration properties

  • transcript (str | None) – Transcription text

  • is_input (bool | None) – Whether transcript is input (True) or output (False)

  • context_messages (str | None) – JSON-serialized context messages

  • function_calls (List[Dict] | None) – Function calls being made

  • tools (List[Dict] | None) – Available tools/functions list

  • tools_serialized (str | None) – JSON-serialized tools for detailed inspection

  • audio_data_size (int | None) – Size of audio data in bytes

  • **kwargs – Additional attributes to add

Return type:

None