Controllable Transcription. Verbatim ( every, filler, pause, stutter, vocal sound) , or intended ( what the speaker meant to say, optimized for readability) with word-level timestamps.
-
Updated
Nov 11, 2026 - Python
Controllable Transcription. Verbatim ( every, filler, pause, stutter, vocal sound) , or intended ( what the speaker meant to say, optimized for readability) with word-level timestamps.
Speakrail - a low-latency fully-local voice assistant
Full-duplex voice AI assistant you can interrupt mid-sentence. Push-to-talk or barge-in over your phone. Local Whisper STT + Piper/Kokoro TTS, runs on your own hardware over a VPN (Tailscale). Self-hosted open-source Python/FastAPI.
Duplex: Real-time full-duplex voice and generative avatar interactions.
Full-duplex I2S audio for the FreeWili 2 (RP2350B / NAU88C10): simultaneous speaker output and external-mic capture, plus 3.5mm TRRS jack output. On-device + webcam/mic E2E validated.
Duplex: Real-time full-duplex voice and generative avatar interactions.
Full-duplex real-time voice debate and clinical persona simulator using React 19 AudioWorklet, FastAPI, and the Gemini Multimodal Live API.
Full-duplex realtime voice bot: Pipecat + Qwen Audio Realtime for ASR/VAD/TTS, Business OpenAPI as the business brain—low-latency browser PCM over WebSocket.
To associate your repository with the full-duplex-audio topic, visit your repo's landing page and select "manage topics."