LoudKit: open-source on-device TTS with voice cloning, 10 languages, 5 SDKs, Apache 2.0
RevolutionaryBox2980 · reddit · 2026-09-10
A developer open-sourced LoudKit, an on-device TTS toolkit built for a reading app, with Apache 2.0 code and weights:
- Capabilities: 10 languages, voice cloning, quality close to Chatterbox (its optimization baseline), small and fast enough to run on an iPhone 14 Pro
- Two variants: loudr-1 and loudr-1-turbo — turbo adds a head that predicts two audio tokens at once on the pipeline's bottleneck, with occasional artifacts
- SDKs in Python, Swift, Go, Rust, TypeScript, with the contract that all backends emit identical speech tokens; audio samples and timestamp metadata may differ slightly
- Cloning tip: 10 seconds of clean recording works best; long pauses or background noise degrade quality. All bundled voices come from consented donations or CC0/CC-BY recordings
- An emotion-tag axis is in progress; the motivation was that existing local TTS attachable to agents like Hermes or OpenClaw still lacks quality
More from coding & agent
- Agent verified 'success' after double-voting: a lesson in harness vs model bugs — TejasKumar_ · 2026-10-26
- diginomica: agentic automation hype is overdone, and 'The Hackening' of enterprise security has begun — jonerp · 2026-10-26
- 'Vibe verification': ~60% of AI-written formal specs prove the wrong property — basedjensen · 2026-10-26
- Devin launches Security suite powered by Claude Mythos with $100M+ credits — DevinAI · 2026-10-26
- HackBench injects reward hacks into agent trajectories to stress-test monitors — DanielKhashabi · 2026-10-26
- Hugging Face ships hf CLI updates: deduped cache saves 19GB, faster buckets, retryable GPU jobs — yb2698 · 2026-10-26