Overview
Miso Labs builds advanced AI for voice applications. It develops foundation models that developers can use to create voice agents with warmth and personality, moving beyond purely functional chat. The core product is a set of voice-focused foundation models that deliver real-time responses with 110-millisecond latency, enabling natural, fluid conversations. It also offers one-shot voice cloning, letting a specific voice be replicated from a 10-second audio sample. For privacy-conscious enterprises, Miso Labs provides open-source models that can run on-premises, so voice data stays in-house under support contracts. Compared with others, its key differentiators are its emphasis on emotive, human-sounding voice, ultra-low latency, one-shot voice cloning capability, and on-prem/open-source options. The company’s goal is to help developers build voice agents that users perceive as more human, friendly, and engaging, with fast, private, and expressive AI interactions.
Work here?Claim your company
Simplify's take
What believers are saying
- Open-weight release on Hugging Face lowers adoption friction for developers.
- YC-backed launch and $8M-plus funding support continued model training.
- On-prem contracts attract regulated buyers keeping voice data in-house.
What critics are saying
- ElevenLabs, Cartesia, Bland, and Hume crush Miso on distribution and brand.
- Latency claims remain unverified by third parties, weakening buyer trust.
- Voice-cloning regulation or copyright litigation could shut down enterprise adoption.
What makes Miso Labs unique
- June 2026 MisoTTS 8B ships open weights and on-prem deployment.
- Miso claims 110ms latency and ten-second one-shot voice cloning.
- Miso One launched June 4, 2026 for emotive, human-like voice agents.
Help us improve and share your feedback! Did you find this helpful?
Funding
Industry standards