AI · AI model
Miso-TTS
Open-source emotive text-to-speech foundation model with low conversational latency.
Miso-TTS is an open-source 8B parameter speech synthesis foundation model developed by Miso Labs. It generates expressive, human-like voice with a fast 110ms response latency and supports one-shot voice cloning. The model can be run locally on-premises or accessed via an API, allowing developers to build real-time voice agents without third-party data exposure.
What it's for
- Build real-time conversational voice agents for customer support or assistants
- Clone a speaker's voice using a 10-second reference audio sample
- Deploy a low-latency speech synthesis model locally for strict data privacy requirements
Features
- 8B parameter architecture built on Llama 3.2 and Mimi audio tokenizer
- 110ms response latency for natural, uninterrupted conversational flow
- One-shot voice cloning from a 10-second audio sample
- Emotive and expressive vocal inflections
- Available for local, on-premises deployment as well as cloud API access
Pricing
Open source. Open-source model download; API access and enterprise support contracts available
Details
- Type
- AI model
- Category
- AI
- Topic
- Voice AI
- Open source
- Yes
- Platforms
- Self-hosted, API, Linux
- For
- Developers and AI engineers building real-time voice applications and voice agents
- Alternative to
- ElevenLabs, Cartesia, PlayHT
- Tags
- tts, speech synthesis, voice cloning, voice ai, real time, open source
Why it was saved
- @girsta: Allows local self-hosting of natural, emotionally varied voice generation and one-shot voice cloning without API fees or audio data privacy concerns.
Compare
Related
Platform for building and deploying interactive voice AI agents.
Open-source Python framework for building realtime voice and multimodal AI assistants
A low-cost, open-source stack for building real-time AI voice assistants
Platform and SDK for building custom AI voice agents and smart devices
No-code platform to build, test, and deploy conversational voice AI agents
Ultra-low-latency streaming text-to-speech and instant voice cloning model.
Updated