Step-by-Step Guide: Integrate DeepSeek AI into Personal WeChat with Docker

Hello, I am a Python advanced user.

Recently DeepSeek has become popular not only because it is open‑source and inexpensive, but also because its writing, coding, and search capabilities are very strong, making it the first open‑source AI in China.

The article provides a hands‑on tutorial for integrating DeepSeek into personal WeChat.

1. Obtain DeepSeek API Key

Visit the DeepSeek developer platform, register if necessary, click “Create API key”, name it, and copy the key. The key can only be copied once; if lost, you must delete and recreate it.

Official usage page: https://platform.deepseek.com/usage

If the DeepSeek website is inaccessible, you can use the AI_API proxy https://kg-api.cloud/ , which also provides a DeepSeek API (paid).

2. Install Docker

Download Docker Desktop for your OS (Apple Silicon, Intel, or Windows) and install it.

For macOS, drag Docker.dmg into the Applications folder.

Then launch Docker Desktop:

Docker.app

3. Configure Docker Image Acceleration (optional)

Reply with “deepseek” to the author’s public account to receive domestic mirror URLs.

4. Create and Configure docker-compose.yml

In any directory, create docker-compose.yml (or use vim on macOS). The file should include the DeepSeek API key and model configuration. Example snippet:

volumes:
  - ./config.json:/app/plugins/config.json

The configuration file config.json contains fields such as "model", "open_ai_api_key", "single_chat_prefix", "group_name_white_list", etc.

{
  "model": "gpt-3.5-turbo",
  "open_ai_api_key": "YOUR API KEY",
  "single_chat_prefix": ["bot", "@bot"],
  "single_chat_reply_prefix": "[bot] ",
  "group_name_white_list": ["ChatGPT测试群"],
  "conversation_max_tokens": 1000,
  "character_desc": "You are an AI assistant based on a large language model..."
}

5. Deploy the Cow Application

Navigate to the directory containing docker-compose.yml and run:

# Check Docker Compose version
docker compose version
# Deploy (v2 syntax)
docker-compose up -d
# or for v3
docker compose up -d

Wait for the image (≈1 GB) to download. When the logs show “started”, deployment is complete.

6. Retrieve WeChat Login QR Code

Run the following command to view the QR code in the Docker logs: docker logs -f chatgpt-on-wechat Or view it in Docker Desktop’s UI.

After scanning the QR code with WeChat, DeepSeek is fully integrated and ready for personal chat.