FastAPI wrapper for Faster-Whisper

Versions

• 1.1.0
  • add v2 router with simple text response
  • add asyncio lock for queening requests
  • default router are deprecated
• 1.0.0
  • /transcribe router

Requirements

• Python3.8 +
• 1.4GB for docker image
• 500MB+ RAM for docker

Installation

1. Set variables. Create .env file with variables:

   MODEL_SIZE="tiny"
   PORT="8080"

2. Run container Simply use docker-compose for run container

   docker-compose up -d

3. Wait for some time while whisper model are downloading.

Usage

Endpoints

• http://127.0.0.1:8080/v2/transcribe - post audio endpoint
• http://127.0.0.1:8080/transcribe - deprecated - post audio endpoint
• http://127.0.0.1:8080/health - healthcheck endpoint
• http://127.0.0.1:8080/docs - fastapi auto documentation

examples

post example

curl -X 'POST' \
  'http://127.0.0.1:8080/v2/transcribe' \
  -H 'accept: application/json' \
  -H 'Content-Type: multipart/form-data' \
  -F 'audio=@voice.ogg;type=video/ogg'

response example

"transcribed text"