LiveKit Adds Fish Audio TTS

// 1h agoPRODUCT UPDATE

LiveKit Adds Fish Audio TTS

LiveKit Agents now supports Fish Audio’s expressive text-to-speech through LiveKit Inference and an official plugin, giving developers low-latency voices for real-time AI agents. The integration includes S2.1 Pro, voice selection, streaming, and expressive delivery controls.

// ANALYSIS

This is a meaningful infrastructure upgrade for voice-agent builders: Fish Audio brings expressive, characterful speech while LiveKit handles real-time orchestration.

–LiveKit Inference removes the need for a separate Fish Audio API key and centralizes usage management.
–Developers can use Fish Audio’s S2 and S2.1 Pro models with configurable latency, speed, volume, temperature, and voice IDs.
–Expressive mode adds emotion, pacing, pauses, laughs, and other non-verbal sounds automatically.
–Streaming synthesis makes the integration suitable for conversational agents, games, companions, and interactive Halloween experiences.
–Custom and community voices still require the direct Fish Audio plugin rather than LiveKit Inference.

// TAGS

livekitfish-audiottsvoice-agentspeechstreamingapi

DISCOVERED

1h ago

2026-10-19

PUBLISHED

1h ago

2026-10-19

RELEVANCE

8/ 10

AUTHOR

FishAudio

// KEEP READING

More AI developer news from the feed

EXPLORE FULL FEED

UPDATE31m ago

interior.dev Adds Pro Tier

interior.dev is adding a Pro tier next week to its free, MIT-licensed React micro-interaction library, which offers 54 copy-paste components built with Tailwind CSS and Motion. The team is asking developers to choose the next interaction to ship.

OPEN SOURCE35m ago

React Three Fiber v10 Alpha 6 Goes WebGPU-First

React Three Fiber v10 alpha 6 makes WebGPURenderer the default and adds deeper Three.js Shading Language support after more than 200 commits. Existing GLSL projects can use the new `/legacy` entry point, while new builds get a WebGPU-first foundation with WebGL2 fallback.

RESEARCH1h ago

NCM Shows 31× Speedup Over Diffusion

Stanford’s Hazy Research presents NCM, a coarse-to-fine framework that predicts hierarchical text codes in parallel, reporting 10–31× faster generation than selected diffusion baselines. Its MIT-licensed [code](https://github.com/HazyResearch/ncm) and checkpoints support reproduction, though evaluations remain limited to small academic benchmarks.[Hazy Research](https://hazyresearch.stanford.edu/blog/2026-10-12-ncm)