English | 简体中文
Article Processor is a tool for article content reproduction and processing. It can automatically handles article formatting, localizes image resources, and use AI to generate summaries, categories and tags.
- Automatically extracts article title, content and source information
- Intelligently optimizes article formatting (including image centering, style standardization, and redundant tag cleanup)
- Standardizes article citations and attribution information
- Downloads and localizes remote image resources
- Automatically converts WebP images to the widely-supported JPEG format
- Centers images for better presentation
- Automatically generates article abstracts
- Generates content summaries
- Creates relevant article tags
- OpenAI
- Google AI Studio
- Custom API endpoints
| Endpoint | Method | Description |
|---|---|---|
| /process | POST | Process an array of article URLs and return JSON results |
| /processStatus/:jobId | GET | Get job status and results |
| /downloadImage:imageName | GET | Retrieve processed images |
git clone https://github.com/elowenluo/article-processor.git
cd article-processor
npm install- Copy the environment template file
cp .env.example .env- Configure required environment variables
# Server access address for generating image URLs
HOST=http://localhost
# Service port
PORT=3000
# Maximum number of concurrent processing tasks (within a job)
MAX_CONCURRENT_TASKS=3
# Maximum number of concurrent jobs (defaults to 1)
MAX_CONCURRENT_JOBS=1# Development environment
npm run dev
# Production environment
npm run build
npm startYou can also run Article Processor using Docker:
docker pull sisuad/article-processor:latestdocker run -d \
-p 3000:3000 \
-e HOST=http://localhost \
-e PORT=3000 \
-e MAX_CONCURRENT_TASKS=3 \
-e MAX_CONCURRENT_JOBS=1 \
-e TZ=Asia/Shanghai \
-v $(pwd)/images:/app/images \
--restart unless-stopped \
--name article-processor \
sisuad/article-processor:latestFor easier deployment, you can use Docker Compose:
docker-compose up -dTo stop the service:
docker-compose downSend a POST request to the /process endpoint with an array of article URLs and AI configuration:
curl -X POST http://localhost:3000/process \
-H "Content-Type: application/json" \
-d '{
"urls": [
"https://news.mydrivers.com/1/1030/1030243.htm"
],
"llmApiConfig": {
"model": "gpt-3.5-turbo",
"apiKey": "sk-XXX"
}
}'curl -X POST http://localhost:3000/process \
-H "Content-Type: application/json" \
-d '{
"urls": [
"https://news.mydrivers.com/1/1030/1030243.htm"
],
"llmApiConfig": {
"model": "gemini-2.0-flash",
"apiKey": "XXX"
}
}'curl -X POST http://localhost:3000/process \
-H "Content-Type: application/json" \
-d '{
"urls": [
"https://news.mydrivers.com/1/1030/1030243.htm"
],
"llmApiConfig": {
"model": "gpt-3.5-turbo",
"url": "https://api.openai.com/v1/chat/completions",
"apiKey": "sk-XXX"
}
}'{
"jobId": "ae587e58-350f-4303-8c59-9c67b09ec189",
"status": "processing",
"message": "Processing started. Check status at the URL below.",
"statusUrl": "/processStatus/ae587e58-350f-4303-8c59-9c67b09ec189"
}Send a GET request to the /processStatus/:jobId endpoint to retrieve the job status and results:
curl -X GET http://localhost:3000/processStatus/ae587e58-350f-4303-8c59-9c67b09ec189{
"jobId": "ae587e58-350f-4303-8c59-9c67b09ec189",
"status": "processing",
"message": "Job is still processing",
"createdAt": "2025-02-27T01:21:06.662Z",
"updatedAt": "2025-02-27T01:21:06.662Z"
}[
{
"title": "DeepSeek成最快突破3000万日活应用程序：手机终端厂商抢着接入DeepSeek",
"content": "<article><p>近日消息，据报道，近期DeepSeek访问使用量急速上升，<span><strong>已经成为目前最快突破3000万日活跃用户量的应用程序。</strong></span></p><p>与此同时，<strong>三家基础电信企业已全面接入国产开源大模型DeepSeek，手机、PC等终端厂商也在积极拥抱DeepSeek，</strong>一些地方政府也开始在政务系统部署DeepSeek。</p><p>比如深圳龙岗区政务服务和数据管理局已经在上线了Deepseek-R1全尺寸模型，成为广东首个在政务信创环境下部署该模型的政府部门单位。</p><p>业内人士表示，DeepSeek的开源模式大幅降低了人工智能进入各行业的门槛，相关的政府和企业级应用有望出现裂变式增长。</p><p>截至目前，在国内市场，DeepSeek直接刷新了豆包、Kimi和文心一言等国内大模型的用户量榜单，2025年1月，DeepSeek月均活跃用户数跃居第一。</p><p>据了解，DeepSeek以3%的成本做出了接近ChatGPT o1水平的模型。低成本便可调校出足够好的AI模型，也让技术闭源的OpenAI和用昂贵算力及CUDA生态拉高壁垒的英伟达神话不攻自破，DeepSeek的成功，让硅谷高管对算力不计成本的投入，一度集体遭到了投资者的质疑。</p><p><figure style=\"text-align: center;\"><img alt=\"DeepSeek成最快突破3000万日活应用程序：手机终端厂商抢着接入DeepSeek\" src=\"https://img1.mydrivers.com/img/20250214/s_a7bd92821ee94f4b9a6abc750efc604f.jpg\"></figure></p> <footer>自 快科技</footer></article>",
"summary": "据报道，DeepSeek成为2025年最快突破3000万日活跃用户的应用程序，并刷新了国内大模型用户量榜单，于2025年1月跃居月均活跃用户数第一。三家基础电信企业、手机及PC终端厂商均已接入DeepSeek。例如，深圳龙岗区政务服务和数据管理局已在政务系统上线Deepseek-R1全尺寸模型，成为广东首个部署该模型的政府部门。DeepSeek以3%的成本实现了接近ChatGPT o1水平的模型效果，其开源模式或将推动政府和企业级应用的增长。",
"tags": ["DeepSeek", "开源大模型", "人工智能", "AI应用", "用户增长"],
"categories": ["人工智能", "数据智能", "移动应用"]
}
]{
"jobId": "ae587e58-350f-4303-8c59-9c67b09ec189",
"status": "failed",
"error": "Failed to process article: Invalid URL or unsupported website",
"createdAt": "2025-02-27T01:21:06.662Z",
"updatedAt": "2025-02-27T01:21:06.662Z"
}- Remove Google Ads
- Support more content source websites
This project is licensed under the MIT License - see the LICENSE file for details.