AdoptAutomate

Kokoro TTS

High-quality local TTS. Best naturalness in the open-source space.

Kokoro produces some of the most natural-sounding speech available from an open-source TTS engine. The prosody and intonation are noticeably better than older models, approaching cloud TTS quality.

The trade-off is speed. Kokoro is slower than Piper, making it better suited for pre-generated audio than real-time streaming. We use it alongside Piper — Kokoro for high-quality cached output, Piper for low-latency responses. The combination covers both use cases.

aittslocal

AdoptPiper TTSFast local text-to-speech. Low latency, many voices.AdoptDiana TTSRust-based TTS engine. Fast inference, small footprint.AdoptOllamaRun LLMs locally. Essential for RAG and embedding pipelines.AdoptLM StudioLocal model server with an OpenAI-compatible API. Drop-in replacement.Adoptnomic-embed-textLocal embedding model. Fast, free, good enough for production RAG.

Kokoro TTS

Related