Open Source Text-to-Speech Projects
Text-to-speech (TTS) technology converts written text into spoken audio. It can make digital content more accessible, support hands-free use, and provide spoken output for applications such as virtual assistants, language-learning tools, and narration. Modern systems can produce natural-sounding speech, vary tone and pacing, and work across different languages and voices. Some run locally, while others use hosted services or APIs.
Open source options include speech synthesis models, voice-cloning and conversion tools, offline libraries, and interfaces for integrating speech into applications. When choosing one, consider output quality, language and voice support, hardware and runtime requirements, license terms, maintenance activity, and compatibility with your existing workflow. These tools are useful to developers, researchers, accessibility teams, and creators who need to generate or customize spoken audio.
23 repositories · updated October 13, 2026
voicebox: Create and Use AI Voices Locally
Voicebox is a desktop voice studio for local speech generation, voice cloning, dictation, and agent audio. It suits creators and developers who want voice tools and integrations on their own machine, with hardware and platform capabilities varying by setup.
ActiveHealth 67/100
Open-LLM-VTuber: Talk to LLMs with a Live2D Avatar
Open-LLM-VTuber is a cross-platform AI companion for voice conversations with language models, paired with a customizable Live2D avatar. It supports local and hosted models, with web and desktop modes for users who want a more expressive, hands-free interface.
QuietHealth 28/100
VoxCPM: Generate and Clone Multilingual Speech
VoxCPM is a tokenizer-free text-to-speech system for multilingual speech generation, voice design, and voice cloning. Its VoxCPM2 release targets teams and developers who need expressive speech synthesis and can support a 2B-parameter model.
ActiveHealth 57/100
MOSS-TTS: Generate Speech, Dialogue, and Sound with AI
MOSS-TTS is a family of speech and sound generation models for long-form narration, voice cloning, dialogue, voice design, sound effects, and streaming TTS. It offers multiple model architectures and deployment paths for research, production, and local inference.
MaintainedHealth 43/100
tortoise.cpp: Generate Speech with Tortoise-TTS in C++
tortoise.cpp is a C++/GGML implementation of Tortoise-TTS for generating speech from text. It can run on CPU or CUDA, but has seen no commits or releases in over a year, so it may suit experiments better than new production deployments.
InactiveHealth 34/100
StreamingKokoroJS: Generate Speech Locally in Your Browser
A browser-based app that turns text into streaming speech with the Kokoro model, keeping processing on your device. It offers WebGPU acceleration with a WebAssembly fallback, but its year-plus without commits makes long-term maintenance uncertain.
InactiveHealth 20/100
chatterbox: Generate Speech from Text with Voice Cloning
Chatterbox is Resemble AI’s Python text-to-speech model family, with English and multilingual options, voice cloning, and expressive speech controls. It suits developers building voice experiences who can manage model inference and its compute requirements.
MaintainedHealth 47/100
Spark-TTS: Generate Speech and Clone Voices from Text
Spark-TTS is a PyTorch inference project for bilingual text-to-speech and zero-shot voice cloning. It uses a Qwen2.5-based model to generate speech and supports adjustable voice characteristics.
InactiveHealth 25/100
RedditVideoMakerBot: Generate Videos from Reddit Posts
RedditVideoMakerBot automates the creation of short videos from Reddit content, without manual video editing or asset compiling. It is for creators who want to generate a video file for manual upload to social platforms.
QuietHealth 39/100
index-tts-lora: Fine-Tune IndexTTS for Custom Voices
A Python project that adds LoRA fine-tuning workflows to IndexTTS for single- and multi-speaker voice synthesis. It is aimed at users who want to adapt speech generation to speaker audio and improve prosody and naturalness.
QuietHealth 20/100
sherpa-onnx: Run Speech AI Locally Across Platforms
sherpa-onnx runs speech and audio models locally through ONNX Runtime, without an internet connection. It supports tasks from speech recognition and synthesis to diarization, with APIs and deployment options spanning mobile, desktop, web, and embedded devices.
ActiveHealth 77/100
ttsfm: Convert Text to Speech with an OpenAI-Compatible API
TTSFM provides a Python client, REST API, and web playground for text-to-speech through a service modeled on OpenAI’s interface. It can suit prototypes and self-hosted integrations, but it relies on a reverse-engineered backend and has had no commits in the last 90 days.
QuietHealth 28/100