pipecat-pyai

PyAI Hear (speech-to-text) and Speak (text-to-speech) services for Pipecat, maintained by PyAI.

Use this package when you want to own the Pipecat transport, frame pipeline, LLM, tools, and turn policy. Use PyAI Omni when you want one managed speech-to-speech agent instead.

pip install pipecat-pyai==0.1.1

Requires Python 3.11 or newer.

Usage

from pipecat_pyai import PyAISTTService, PyAITTSService

stt = PyAISTTService()             # PyAI Hear (streaming STT)
tts = PyAITTSService(voice="stock_emma_en_gb")  # PyAI Speak (TTS)

Set your key once: export PYAI_API_KEY=pyai_live_... (or pass api_key=). The key needs the hear:stream, speak:synthesize scopes. List voices at GET https://api.pyai.com/v1/voices. See the integration status and comparison for the current public guide link, production notes, and release status. The source distribution also includes examples/bot.py.

PyAISTTService

WebSocket streaming STT (subclass of Pipecat's WebsocketSTTService). Sends pipeline audio to the PyAI Hear socket and pushes InterimTranscriptionFrame (PyAI partial/speech_final) and TranscriptionFrame (PyAI corrected final). Args: api_key, base_url, model (pyai-hear), language, sample_rate. The service sends the canonical commit control when Pipecat's VAD reports the end of speech and marks the corrected transcript finalized.

PyAITTSService

Chunked TTS (subclass of TTSService). Requests response_format: pcm at the pipeline's output sample rate and pushes TTSAudioRawFrames as the audio streams. Args: voice (required), api_key, base_url, model (pyai-speak), sample_rate, aiohttp_session. Use a streaming-capable voice id.

Notes

• Billing/limits follow your PyAI key (hear.minutes, speak.minutes); a streaming STT session counts against your key's realtime concurrency.
• Built on the official PyAI Python SDK (pyai-sdk).
• Hear streaming supports en, es, fr, de, hi, it, pt, and nl. Pass the explicit language for non-English audio.
• Pipecat Cloud, transports, your LLM, and worker hosting are separate from PyAI Hear and Speak metering.
• For a fully managed agent, including turn-taking, reasoning, tools, caller continuity when enabled on an Agent profile and a caller key is available, and optional telephony, use PyAI Omni instead.
• Licensed under Apache-2.0. The source is included in the wheel and source distribution.

Language is detected automatically when language is omitted. Pass an explicit supported language only when you intend to pin recognition.