How to Connect DeepSeek AI to Your Personal WeChat Using Docker – Step‑by‑Step Guide

DeepSeek, China’s first open‑source AI model, offers strong writing, coding, and search capabilities. This guide shows how to integrate DeepSeek into a personal WeChat account using the open‑source chatgpt‑on‑wechat (cow) project.

1. Obtain a DeepSeek API Key

Visit the DeepSeek platform at https://platform.deepseek.com/usage to create an account and generate an API key. If the official site is inaccessible, use the proxy service https://kg-api.cloud/ as an alternative endpoint.

Official API endpoint: https://api.deepseek.com/v1 Proxy endpoint: https://kg-api.cloud/v1

2. Install Docker

Download Docker Desktop for your operating system:

macOS – choose the Apple Silicon or Intel package.

Windows – download from the Docker website.

After installation, launch Docker Desktop.

3. Create and Configure docker-compose.yml

In a directory of your choice, create a docker-compose.yml file (or edit the existing one). Include the following basic structure and replace placeholders with your own values:

version: '3'
services:
  chatgpt-on-wechat:
    image: ghcr.io/zhayujie/chatgpt-on-wechat:latest
    container_name: chatgpt-on-wechat
    restart: unless-stopped
    ports:
      - "8080:8080"
    volumes:
      - ./config.json:/app/plugins/config.json
    environment:
      - OPENAI_API_KEY=YOUR_API_KEY
      - OPENAI_API_BASE=https://api.deepseek.com/v1

A sample config.json (placed beside the compose file) can contain:

{
  "model": "gpt-3.5-turbo",
  "open_ai_api_key": "YOUR_API_KEY",
  "single_chat_prefix": ["bot", "@bot"],
  "single_chat_reply_prefix": "[bot] ",
  "group_chat_prefix": ["@bot"],
  "group_name_white_list": ["ChatGPT测试群"],
  "speech_recognition": false
}

4. Deploy the Cow Application

Open a terminal in the directory containing docker-compose.yml and run:

# Check Docker Compose version
docker compose version
# Deploy in detached mode (v2 syntax)
docker compose up -d
# Or for v3 syntax
# docker-compose up -d

The image download may take a few minutes. When the logs show a started message, the service is running.

5. Retrieve the WeChat QR Code

Execute the following command to view the QR code for WeChat login: docker logs -f chatgpt-on-wechat The QR code will appear in the terminal or in Docker Desktop’s log view. Scan it with WeChat to complete the connection.

6. Test the Installation

Verify Docker works:

docker --version
docker run hello-world

If the hello-world container runs without error, Docker is ready for the cow service.

After completing these steps, DeepSeek is fully integrated into your personal WeChat, allowing private and group chats, voice recognition, image generation, and more.

Images illustrating the process: