Open Source Models for Voice AI

OpenHome — OpenHome — open-source OS for AI voice agents, LLM-agnostic and ship-ready for real hardware.

QwenLM — Qwen3-ASR is an open-source series of ASR models developed by the Qwen team at Alibaba Cloud, supporting stable multilingual speech/music/song recognition, language detection and timestamp prediction.

DeepSpeech — DeepSpeech is an open source embedded (offline, on-device) speech-to-text engine which can run in real time on devices ranging from a Raspberry Pi 4 to high power GPU servers.

Dograh — Open-source voice platform for custom, self-hosted voice agents and pipelines.

Zonos — Zonos-v0.1 is a leading open-weight text-to-speech model trained on more than 200k hours of varied multilingual speech, delivering expressiveness and quality on par with—or even surpassing—top TTS providers.

Voicebox — Open Source Voice Cloning Desktop App for Mac, Windows, and Linux.

Common Voice — Mozilla's open, crowdsourced multilingual speech dataset for training speech recognition — millions of validated voice clips with transcripts across 100+ languages, released under CC0. The platform code is MPL-2.0.

FluidVoice — FluidVoice is an open-source voice dictation / speech-to-text tool.

Ultravox — Real-time, speech native voice AI infrastructure for fast, natural, scalable voice agents.

F5-TTS — Official code for "F5-TTS: A Fairytaler that Fakes Fluent and Faithful Speech with Flow Matching"

FireRedTeam — FireRedChat is FireRedTeam's open-source real-time, full-duplex conversational voice agent.

Vosk — Offline speech recognition API for Android, iOS, Raspberry Pi and servers with Python, Java, C# and Node

kokoro — https://hf.co/hexgrad/Kokoro-82M

Gemma — Gemma open-weight LLM library, from Google DeepMind

VoiceBun — Voice assistant brand focused on precise descriptions to improve results.

LobbyStack — Open-source AI receptionist that answers calls, qualifies leads, books appointments, and routes urgent requests 24/7.

Coqui TTS — 🐸💬 - a deep learning toolkit for Text-to-Speech, battle-tested in research and production

TEN Framework — Voice Activity Detector (VAD) : low-latency, high-performance and lightweight

Kyutai — Pocket TTS is a compact text-to-speech model small enough to run on a CPU — fast, lightweight, on-device voice generation.

ChatTTS — A generative speech model for daily dialogue.

KugelAudio — Europe's first production-ready TTS with 40+ languages, developed and hosted in Europe, fully GDPR compliant.

Moonshot AI — MoBA: Mixture of Block Attention for Long-Context LLMs

SpeechBrain — A PyTorch-based Speech Toolkit

Whisper — Robust Speech Recognition via Large-Scale Weak Supervision

Rasa — The developer platform for enterprise AI agents.

Pebble — Pebble Index 01: private on-device memory ring to capture ideas.

neutts — On-device TTS model by Neuphonic

Moonshine — Very low latency speech to text, intent recognition, and text to speech, for building voice agents and interfaces

GLM-5 — GLM-5: From Vibe Coding to Agentic Engineering

fish-speech — SOTA Open Source TTS