DDG Chat

Deploy your ChatGPT API backend to various platforms for free with one click

Supports Vercel, Cloudflare Workers, Docker, Render, etc.

Support GPT4o mini, Claude 3 Haiku, Llama 3.1 70B, Mixtral 8x7B models

All models provided anonymously by DuckDuckGo

One-click deployment

Manual deployment

Vercel / Cloudflare Workers / Docker / Render / Hugging Face / More Serverless

Demo API

Do not abuse

• https://chatcfapi.r12.top

Recommended usage

Use a third-party ChatGPT application to call the interface, for example:

• ChatNextWeb
• immersive translation

Call interface

Conversation request

Example (please replace chatcfapi.r12.top with your own deployed domain name):

curl --request POST ' https://chatcfapi.r12.top/v1/chat/completions ' 
  --header ' Content-Type: application/json ' 
  --data ' {
    "messages": [
      {
        "role": "user",
        "content": "你好！"
      }
    ],
    "model": "gpt-4o-mini",
    "stream": true
  } '

Model query

Example (please replace chatcfapi.r12.top with your own deployed domain name):

Visit http://chatcfapi.r12.top/v1/models to view the currently supported models

Supported models

The model name is obtained from DDG web page capture simultaneously (unknown models will be redirected to the gpt-4o-mini model)

• gpt-4o-mini
• claude-3-haiku
• llama-3.1-70b
• mixtral-8x7b

Manual deployment

Since the DDG API limits the number of concurrencies for a single IP, it is recommended to use Vercel for deployment. If you use local deployment such as Docker, please ensure that the project runs in the agent pool.

Vercel

Method 1: Cloud Fork warehouse deployment

1. Fork this repository to your Github
2. Enter the Vercel New Project web page and Import the warehouse you just forked.
3. Click Deploy

Method 2: Local Clone warehouse deployment

1. First make sure you have a Nodejs environment
2. Execute the following command:

npm i -g vercel
vercel login
git clone https://github.com/leafmoes/DDG-Chat.git ddg-chat
cd ddg-chat
npm run publish

Method 3: One-click deployment

Render

1. Fork this repository to your Github
2. Click the button below to go to the Render official website, Import the warehouse you just forked, and then follow the normal deployment process.

Cloudflare Workers

Method one:

1. Go to the Cloudflare Workers and Pages console and create a Workers
2. Enter the Workers settings and set兼容性标志to nodejs_compat at运行时
3. Paste the code inside this file into your Workers and click Deploy

Method two:

1. First make sure you have a Nodejs environment
2. First execute the following command:

npm i -g wrangler
wrangler login
git clone https://github.com/leafmoes/DDG-Chat.git ddg-chat
cd ddg-chat
npm install

3. Open the file /api/index.js, uncomment the last line, and then execute npm run publish:cf

Hugging Face

1. Create a new space: Create a new Space

    Space name：ddg-chat
   License： MIT
   Select the Space SDK： docker （Blank）
   Space hardware： free
   

2. To set environment variables, go to Settings -> Variables and secrets -> Variables. It seems that API_PREFIX must be set to configure environment variables. Due to the restriction of hugging face? Environment variables are in the post

3. Edit the README.md file and add app_port: 8787 in the metadata. This 8787 depends on the port of your docker image.

4. Add Dockerfile file:

    FROM docker.io/leafmoes/ddg-chat:latest
   

Docker

Method 1: Command line build

docker run -it -d --name ddg-chat -p 8787:8787 leafmoes/ddg-chat:latest

Method 2: Build using docker-compose.yml file

Download and save the docker-compose.yml file, and then run docker-compose up -d in the directory where the file is located to start the service.

Docker related commands

docker logs -f ddg-chat # 查看服务实时日志
docker restart ddg-chat # 重启服务
docker stop ddg-chat # 停止服务

environment variables

 # API 服务使用的端口
PORT = 8787
# API 调用的前缀地址
API_PREFIX = ' / '
# 作为调用 API 验证的 API Key
API_KEY = ' dummy_key '
# 向 DDG 发送请求失败的重试次数
MAX_RETRY_COUNT = 3
# 向 DDG 发送请求失败的重试延迟，单位 ms
RETRY_DELAY = 5000

FAQ

1. What are the limitations of Vercel deployment?
   • The free version has a single API request duration of 60 seconds.
   • The free version has 100,000 API calls per month
   • Official Documentation - Vercel Function Limitations
2. Are there plans to add front-end applications?
   • If not considered, please use other excellent projects, such as: ChatNextWeb
3. Why is error 429 ERR_SERVICE_UNAVAILABLE reported? How to solve it?
   • Non-Vercel is prone to this problem. Since the DDG API limits single IP concurrency, it is recommended to reduce the amount of concurrency and use a proxy pool for requests.

communication group

Telegram

Contributors

LICENSE

MIT License