Software Mansion ships Reverb-TTS: one-clip voice cloning TTS that runs locally in the browser
Scobleizer · x · 2026-11-20
Software Mansion AI Labs' first model, Reverb-TTS, is a state-of-the-art voice-cloning text-to-speech model for local inference. It clones a voice from one short clip and speaks Polish, German, French, Spanish, Italian, Portuguese, and English on a phone — small enough to run locally in a browser. It claims SOTA on both word error rate and speed, and is available via react-native-executorch and ruNNtime, with a Mac demo through Neywa Labs' Nativ.
Related event: Software Mansion Open-Sources Reverb-TTS for Local Voice Cloning(3 posts)→
More from Infra
- GPU rents up, token prices down: inference providers' break-even load jumps 22%–75% in 5 months — AccBalanced · 2026-11-20
- Missing config file was throttling GPUDirect Storage — fix yields 3x reads, 9x writes — AccBalanced · 2026-11-20
- NVIDIA Dynamo cuts RL rollout bottlenecks with KV-aware routing and delta weight updates — NVIDIAAI · 2026-11-20
- Arm reaches the AI laptop: Surface Laptop Ultra runs 120B models locally — kimmonismus · 2026-11-20
- Agents pull AI compute back to the CPU: AWS, Google, Microsoft, NVIDIA all build on Arm — kimmonismus · 2026-11-20
- Azure Cobalt 200 goes GA at Ignite: Microsoft's 132-core Arm chip built for agent workloads — kimmonismus · 2026-11-20