Kokoro-FastAPI is a text-to-speech API server that wraps the Kokoro-82M model in a Dockerized FastAPI application.
The tool solves the problem of generating high-quality speech from text by providing an OpenAI-compatible API endpoint around Kokoro-82M, a lightweight text-to-speech model. It supports multiple languages including English, Spanish, French, Hindi, Italian, Japanese, Brazilian Portuguese, and Mandarin Chinese. The approach works by containerizing the model with prebuilt Docker images optimized for different hardware platforms, allowing users to deploy it locally without managing dependencies manually. Beyond basic speech generation, the tool offers custom voice creation through clone-tuning, per-word or per-chunk timestamped captions, inline multi-speaker generation with voice mixing, SSML support, and phoneme-level control.
Adoption suits developers building voice features into applications who want a self-hosted alternative to cloud APIs. The project is particularly valuable for those integrating with Home Assistant, building audiobook or ebook readers, creating AI assistants, or developing browser extensions. The prebuilt multiplatform images support CPU inference on both x86_64 and ARM64 architectures, NVIDIA GPUs across multiple CUDA versions including recent Blackwell hardware, experimental AMD ROCm support, and Apple Silicon via native MPS. An optional integrated web UI with read-along functionality is included for long-form generation tasks.
Development shows consistent maintenance with regular updates to support new hardware generations and integration patterns. The project maintains comprehensive Docker image variants across multiple platforms and GPU architectures, indicating active attention to deployment scenarios. Community adoption is evidenced by integration into established platforms like Home Assistant, Umbrel, Unraid, and various specialized applications for reading, audiobooks, and AI assistants. The README documents multiple integration paths and community projects, suggesting the maintainer actively tracks and supports ecosystem growth.
