Media
Text Frames
Frame types for handling text and transcriptions in Pipecat
TextFrame
Base class for all text-based frames. Used for general text content, transcriptions, and LLM interactions.
Properties
text
str
requiredThe text content being carried by the frame
Methods
TranscriptionFrame
Frame for final transcription results from speech-to-text services.
Properties
text
str
requiredThe transcribed text
user_id
str
requiredIdentifier of the user who spoke
timestamp
str
requiredISO 8601 formatted timestamp of when the transcription was generated
language
Language | None
Detected language of the transcription. See Language enum for supported values.
Methods
InterimTranscriptionFrame
Frame for preliminary transcription results during ongoing speech recognition.
Properties
Identical to TranscriptionFrame:
text
str
requiredThe preliminary transcribed text
user_id
str
requiredIdentifier of the user who is speaking
timestamp
str
requiredISO 8601 formatted timestamp
language
Language | None
Detected language of the transcription
Usage Examples
Creating Text Frames
Common Pipeline Usage
Frame Flow
Notes
- InterimTranscriptionFrames are used for real-time feedback
- TranscriptionFrames represent final, confirmed transcriptions
- Language detection is optional and may not be available from all STT services