docker-insanely-fast-whisper-api

Docker container for insanely-fast-whisper with API, UI and multi-user Auth for insanely-fast-whisper

API for whisper AI that supports multiple user authentication. Mostly implemented with Nodejs.

Installation

git clone https://github.com/lojik-ng/docker-insanely-fast-whisper-api.git
cd docker-insanely-fast-whisper-api
mv server/keys.sample.json server/keys.json
docker build -t docker-insanely-fast-whisper-api .
docker run -d -it -p 2901:2901 --gpus all  --restart=unless-stopped -v .:/shared --name docker-insanely-fast-whisper-api docker-insanely-fast-whisper-api

You can now access the ui at http://localhost:2901/.

Endpoints/API

• Send a POST request to http://localhost:2901/transcribe with {audio: file, apiKey: string}
• Check server/index.html for example usage of the endpoints

API Keys/Authentication

• Edit server/keys.json in the cloned repository anytime to add or remove API keys.

Models

Edit server/model.txt to the model you prefer. You can also use distil models.

• openai/whisper-tiny
• openai/whisper-tiny.en
• openai/whisper-base
• openai/whisper-base.en
• openai/whisper-small
• openai/whisper-small.en
• openai/whisper-medium
• openai/whisper-medium.en
• openai/whisper-large
• openai/whisper-large-v3
• etc

Logging

• logs are rotated daily and can be found in /logs folder of the cloned repository.
• Logs are never purged. You'll need to manually purge the logs.
• Access logs, error logs etc are lumped together

Features

• Requires GPU. I tested with Nvidia GPU (Cuda).
• API is written in NodeJS so should be easy for JS devs to modify. Check server/index.js.
• Authentication: You can add as many users and ban users just by editing the server/keys.json file.
• Logging: It logs all user requests, errors etc.

Todo

• Speed up inference with Flash attention 2.
• Create translation API