Kokoro FastAPI OpenAI-Compatible Text-to-Speech Server

Kokoro-FastAPI is a Dockerized FastAPI wrapper around the Kokoro-82M text-to-speech model with OpenAI-compatible speech endpoints. It supports local TTS serving, multi-language synthesis, web UI access, and timestamped audio generation workflows.

Prerequisites

Docker

Installation

Use the upstream install or setup path that matches your environment:

• docker run -p 8880:8880 ghcr.io/remsky/kokoro-fastapi-cpu:latest # CPU, or:
• docker run --gpus all -p 8880:8880 ghcr.io/remsky/kokoro-fastapi-gpu:latest # NVIDIA GPU, or:
• docker run --device=/dev/kfd --device=/dev/dri -p 8880:8880 ghcr.io/remsky/kokoro-fastapi-rocm:latest # AMD GPU (ROCm, experimental, amd64 only)
• git clone https://github.com/remsky/Kokoro-FastAPI.git

Requirements and caveats from upstream:

• <summary>Quickest Start (docker run)</summary>
• <summary>Quick Start (docker compose) </summary>
• Install prerequisites, and start the service using Docker Compose (Full setup including UI):

Basic usage or getting-started notes:

• Pre built images are available to run, with arm/multi-arch support, and baked in models

• Named versions should be pinned for your regular usage.

• Source: https://github.com/remsky/Kokoro-FastAPI

• Extracted from upstream docs: https://raw.githubusercontent.com/remsky/Kokoro-FastAPI/HEAD/README.md

Documentation

• https://github.com/remsky/Kokoro-FastAPI/wiki

Source

• Agent Skill Exchange