Skip to content

Latest commit

 

History

History
162 lines (117 loc) · 4.24 KB

HowToRun_en.md

File metadata and controls

162 lines (117 loc) · 4.24 KB
Raw

Production Version

Step1: Deploy with Docker

cd docker
docker-compose up -d

Access http://127.0.0.1:3000 in browser

Step2: Configure Model API Key, Base URL and Model Name in Settings

This project requires LLM and TTS services. There are multiple options available. Please read the configuration guide carefully 😊

1. Get LLM API_KEY:

Recommended Model Provider base_url Price Effect
gemini-1.5-pro-002 Yunwu API https://yunwu.zeabur.app ¥7 / 1M tokens 🤩
claude-3-5-sonnet-20240620 Yunwu API https://yunwu.zeabur.app ¥10 / 1M tokens 🤩
gpt-4o Yunwu API https://yunwu.zeabur.app ¥7 / 1M tokens 😃

⚠️ Note: Claude, Grok and domestic models like Deepseek are recommended. GPT-4 is not recommended.

How to get API key from Yunwu API?
  1. Go to Yunwu API website
  2. Register an account and top up
  3. Create a new key in the API key page
  4. Make sure to check Unlimited quota and select Pure AZ 1.5x channel
Can I use other models?
  • ✅ OAI-Like API interfaces are supported, you can change it in Settings.
  • ⚠️ However, other models (especially smaller ones) have weak instruction following capabilities and are prone to errors during translation. Strongly not recommended. If you encounter errors, please switch models.

Claude 3.5 Sonnet and Deepseek models work well!

Recommend using API from https://yunwu.zeabur.app/ (can use Claude at a low price, super cheap!)

If you pursue stability, you can use openrouter

Development Version

Requirements

  • Python 3.8+
  • Node.js 16+
  • MongoDB 4.0+

I. Backend Deployment

1. Install Dependencies

cd api
pip install -r requirements.txt

2. Configuration File

Copy the configuration template to create the actual config:

cp config.example.yml config.yml

Edit config.yml:

mongodb:
  uri: "mongodb://username:password@localhost:27017/dbname?authSource=admin"
  # Modify with your MongoDB connection info

secret_key: "your-secret-key-here" 
# Set a random string as secret key

llm_api_key: "your-openai-api-key-here"
# Enter your API key

llm_base_url: "https://api.wlai.vip/v1"  
# Official API URL: https://api.openai.com/v1
# Third-party API URL depends on provider

llm_model: "claude-3-5-sonnet-20241022"  
# Available models: gpt-4-turbo-preview, gpt-3.5-turbo, etc.

3. Start Backend

cd api
python app.py

Defaults to http://localhost:5000

II. Frontend Deployment

1. Install Dependencies

cd web

# Using npm
npm install

# Or using pnpm (recommended, faster)
pnpm install

# Or using cnpm (for poor network conditions in China)
cnpm install

2. Run in Development Environment

npm run dev

Defaults to http://localhost:3000

3. Production Build

npm run build

Build output in dist directory

III. Important Notes

  1. MongoDB Related:

    • Ensure MongoDB service is running
    • Configure database username and password correctly
    • Ensure database port is accessible

  2. API Related:

    • Ensure API Key is valid with sufficient quota
    • If using third-party API, ensure service stability

  3. Network Related:

    • Check backend CORS configuration if encountering cross-origin issues
    • Ensure frontend and backend ports are not occupied

  4. Production Deployment:

    • Recommend using nginx as frontend static resource server
    • Recommend using gunicorn or uwsgi as WSGI server for backend
    • Recommend configuring SSL certificate for HTTPS

IV. Common Troubleshooting

  1. Frontend Cannot Connect to Backend:

    • Check if backend service is running properly
    • Verify frontend API address configuration
    • Check network connection and firewall settings

  2. MongoDB Connection Failed:

    • Check MongoDB service status
    • Verify connection string format
    • Confirm user permission configuration

  3. API Call Failed:

    • Check API Key configuration
    • Confirm model name is correct
    • Check API quota balance