Simple Whisper API

Description

This project implements an audio transcription API using OpenAI's Whisper model and FastAPI. The API provides endpoints for transcribing audio from a provided URL or an uploaded audio file. It allows for choosing between different sizes of Whisper models and configuring transcription parameters.

Features

• Transcribe audio from a URL.
• Transcribe audio from an uploaded file.
• Support for multiple Whisper model sizes (tiny, base, small, medium, large).
• Configurable language and other transcription parameters.

Installation

Before starting, ensure you have Python 3.8+ installed. Follow these steps to install dependencies and run the server locally:

git clone https://github.com/wildsonc/whisper-api.git
cd whisper-api
pip install -r requirements.txt
uvicorn main:app --host 0.0.0.0 --port 80

Usage

After starting the server, the API will be available at http://127.0.0.1:80.

Docs are available at http://127.0.0.1:80/docs.

Endpoints

• POST /transcribe: Transcribes audio from a URL.
• POST /transcribe/file: Transcribes audio from an uploaded file.

Body

{
  "url": "http://example.com/audio.mp3",
  "model": "small", # optional, defaults to "small"
  "settings": { # optional
    "language": "pt",
    "initial_prompt": "Some text",
    "temperature": 0.4
  }
}

File Upload

To transcribe audio from an uploaded file, send a POST request to /transcribe/file. This endpoint accepts form data with the audio file, model selection, and optional settings.

cURL Example

curl -X POST "http://127.0.0.1:80/transcribe/file" \
     -H "accept: application/json" \
     -H "Content-Type: multipart/form-data" \
     -F "file=@path_to_your_audio_file.mp3;type=audio/mpeg" \
     -F "model=small" \
     -F "language=pt" \
     -F "initial_prompt=Some text"