ASR (Automatic Speech Recognition) Service

This project uses Automatic Speech Recognition (ASR) technology to convert speech to text, supporting multiple ASR services.

Standalone Whisper Deployment

If you want to deploy Whisper speech recognition service locally, you can try using the docker image onerahmet/openai-whisper-asr-webservice:latest-gpu and deploy through the following steps:

Environment Requirements

• NVIDIA GPU
• Docker and Docker Compose
• CUDA-enabled server

Deployment Steps

1. Create a docker-compose.yml file with the following content:

version: '3'
services:
  whisper:
    image: onerahmet/openai-whisper-asr-webservice:latest-gpu
    ports:
      - 9000:9000  # Service exposure port
    environment:
      - ASR_MODEL=base
      - DEVICE_TYPE=cuda:0  # If no GPU available, set to cpu
    restart: unless-stopped
    volumes:
      - ./audio_data:/tmp  # Audio file storage directory

2. Run in the directory:

docker-compose up -d

3. Test if the service is running:

curl -X POST "http://localhost:9000/asr" -H "accept: application/json" -H "Content-Type: multipart/form-data" -F "audio_file=@test.wav;type=audio/wav"

Cloud ASR Services

Option 1: Groq API (Recommended for International Users)

• Fast processing speed
• Good accuracy
• Competitive pricing

Option 2: Silicon Flow API (Recommended for Domestic Users)

• Stable domestic access
• Chinese language optimization
• Good cost-effectiveness

Configuration Steps

1. Register account for your chosen service
2. Obtain API Key
3. Configure in Amadeus system settings
4. Test voice recognition functionality

Service Selection

Choose the appropriate ASR service based on your geographic location and language requirements.