huggingface/speech-to-speech

Build local voice agents with open-source models

Python13K stars1.6K forks+393 today

Topics

#ai#assistant#language-model#machine-learning#python#speech#speech-synthesis#speech-to-text#speech-translation

What it does

Speech To Speech is a modular voice-agent pipeline that enables the creation of local voice agents using open-source models, providing low-latency communication through an OpenAI Realtime-compatible WebSocket API. It matters for developers and researchers who want to build voice-based applications without relying on proprietary services.

Star history

Not enough history yet — 1 day(s) recorded. The daily snapshot builds this up.

Tracking

Last trending: 2026-08-01

Creator kit

Hook

Transform your ideas into voice-enabled apps with Speech To Speech, the open-source pipeline for building local voice agents. #VoiceAgents #OpenSourceAI

Content angles

Create a tutorial on setting up a custom voice assistant using this framework.
Develop and share a comparison of different LLMs used in the pipeline.
Demonstrate how to integrate Speech To Speech with existing IoT devices.

Who should care

Developers, researchers, and hobbyists interested in AI-driven voice applications.