GitHub - LearningByDoingNow/ai-hub: 一个AI驱动的可扩展的资讯fetch工具，一站式浏览 AI 公司产品、最新资讯与前沿论文。An AI-powered, extensible news fetching tool — a one-stop hub for browsing AI company products, latest news, and cutting-edge research papers.

AI-Powered Global Information Aggregation Platform
_{Smart fetching, filtering, and aggregation — never miss what matters}

Live Preview^* • Quick Start • Features • Desktop Widget • Tech Docs • User Guide • 中文

_{* Online preview is a simplified read-only demo. For full features (AI chat, settings, auto-fetch, desktop widget), please clone and run locally.}

What is AI Hub?

AI Hub automatically aggregates content from 77+ premium global sources across multiple domains — AI technology, academic papers, international affairs, and more. It features a modern WebUI and a native macOS desktop widget, both powered by the same data engine and sharing a unified database.

Two ways to use:

WebUI — Full-featured web interface with search, filtering, favorites, settings, AI chat
Desktop Widget — Lightweight floating widget with real-time notifications, favorites, and AI assistant

Both share the same database and configuration — favorites, read status, and sources stay in sync.

Features

Intelligent Aggregation Engine

77+ curated data sources — OpenAI, DeepMind, TechCrunch, BBC, Financial Times, arXiv, and more
Smart filtering — AI-relevance detection, 7-day freshness window, duplicate prevention
Parallel fetching — All sources fetched concurrently with 30s deadline, completes in ~8 seconds
Configurable auto-fetch — Set any interval (1min to hours), runs in background

WebUI

Homepage

Dark-themed dashboard with stats overview, latest news, recent papers, and featured AI products.

AI News Aggregation

Real-time AI news from top sources with source-type filtering (Twitter, WeChat, RSS) and search.

Paper Tracker

Track cutting-edge research papers from arXiv (cs.AI, cs.LG, cs.CL, cs.CV) with direct links to papers and PDFs.

World News

International affairs coverage from BBC, Financial Times, NYT, Reuters, Guardian, and more — with source filtering.

AI Products Directory

Browse 59+ AI companies across 9 categories with direct links to official sites, APIs, and documentation.

Favorites

Unified favorites system shared between WebUI and Desktop Widget — save and manage your bookmarks.

Settings & Configuration

Comprehensive settings panel with 5 tabs:

Tab	Description
Data Fetching	One-click fetch, auto-fetch interval, recent fetch history
Module Management	Create/edit content modules (AI News, Papers, World News, etc.)
Data Sources	Manage 77+ RSS sources, assign to modules
Product Management	CRUD for AI products directory
LLM Configuration	Quick-select presets (OpenAI, Anthropic, DeepSeek, GLM, etc.)

Data Fetching Demo

LLM Configuration Demo

Desktop Widget (macOS)

A native floating widget that lives on your desktop — always accessible, never in the way.

Expanded Card List

_{Read status, favorites, source filtering, dismiss cards}

Full Desktop View

_{Widget + AI Chat + Settings side by side}

Widget Features:

Floating logo with particle effects — click to expand
Source-type filter tabs (All, Twitter, WeChat, RSS, World)
Read tracking (persisted across restarts)
Favorites (synced with WebUI)
Dismiss individual cards or clear all
AI Chat assistant with streaming responses
Settings panel with LLM configuration
Drag to reposition, resize by dragging bottom edge

Quick Start

Prerequisites

Requirement	Version	Purpose
Node.js	18+ (recommend 20+)	WebUI + data fetching engine
npm	9+	Package management
Docker	(Optional)	WeChat sources via WeWe RSS
Rust	(Optional)	Build desktop widget from source

1. Clone & Install

git clone https://github.com/LearningByDoingNow/ai-hub.git
cd ai-hub
npm install

2. One Command Start

npm run start:all

This single command automatically:

Starts WeWe RSS Docker container (if Docker is installed, for WeChat sources)
Fetches data from all 77+ sources (first run)
Starts the WebUI at http://localhost:3000
Starts the WeWe watcher (auto-fetches when WeChat has new content)

╔══════════════════════════════════════╗
║        AI Hub — Starting Up          ║
╚══════════════════════════════════════╝

[WeWe] Started (port 4000, scan every 10min)
[Watch] Starting WeWe watcher (auto-fetch on update)
[WebUI] Starting at http://localhost:3000

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
  ✓ AI Hub is running!
    WebUI:     http://localhost:3000
    WeWe RSS:  http://localhost:4000
    Watcher:   Auto-fetch on WeChat updates
    Press Ctrl+C to stop
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

No Docker? No problem. The script auto-detects — without Docker, WeChat sources are skipped but all other 60+ sources (RSS, arXiv, etc.) work perfectly.

3. (Optional) Configure LLM for AI Chat

cp .env.example .env.local

Edit .env.local:

LLM_BASE_URL=https://open.bigmodel.cn/api/paas/v4   # Or any OpenAI-compatible endpoint
LLM_API_KEY=your_api_key_here
LLM_MODEL=glm-4.5-air                                # Model name
LLM_TEMPERATURE=0.5

Or configure directly in WebUI: Settings → LLM Configuration (with quick-select presets).

Open http://localhost:3000 — you'll see all aggregated content immediately.

4. Configure LLM (Optional — for AI Chat)

cp .env.example .env.local

Edit .env.local:

LLM_BASE_URL=https://open.bigmodel.cn/api/paas/v4   # Or any OpenAI-compatible endpoint
LLM_API_KEY=your_api_key_here
LLM_MODEL=glm-4.5-air                                # Model name
LLM_TEMPERATURE=0.5

Supported providers: OpenAI, Anthropic, DeepSeek, GLM (智谱), Together AI, Groq, SiliconFlow, Ollama (local), etc.

You can also configure LLM directly in the WebUI: Settings → LLM Configuration (with quick-select presets).

Available Commands

Data Fetching

Command	Description
`npm run fetch:all`	One-shot: fetch news + papers from all sources (~8s)
`npm run fetch`	One-shot: fetch news only
`npm run fetch:papers`	One-shot: fetch arXiv papers only
`npm run fetch:schedule`	Loop: auto-fetch every 4 hours (runs in foreground)

WebUI

Command	Description
`npm run dev`	Start dev server with hot-reload (http://localhost:3000)
`npm run build`	Production build
`npm run start`	Start production server

Desktop Widget

Command	Description
`npm run desktop:install`	Install desktop dependencies
`npm run desktop:dev`	Start desktop widget in dev mode (hot-reload)
`npm run desktop:build`	Build .app and .dmg for distribution

Quick Launch (Full Stack)

Open 4 terminals for the complete experience:

# Terminal 1: Start WeWe RSS (if using WeChat sources)
docker start wewe-rss

# Terminal 2: Start WebUI
npm run dev

# Terminal 3: Start Desktop Widget
npm run desktop:dev

# Terminal 4: Auto-fetch watcher (detects WeWe updates → fetches immediately)
npm run fetch:watch

Or a minimal single-terminal start:

npm run fetch:all && npm run dev

Auto-fetch Options

Method	How	Best For
`npm run fetch:watch`	Monitors WeWe RSS every 2min, auto-fetches on update	Recommended — real-time WeChat updates
`npm run fetch:watch -- 60`	Same but checks every 60 seconds	Faster detection, slightly more CPU
WebUI	Settings → Data Fetching → Set interval	Periodic fetch of all sources
Desktop Widget	Settings → Set interval	Same as WebUI
`npm run fetch:schedule`	Loop every 4 hours	Background cron-style
System cron	`crontab -e` → `0 /4 * * cd /path/to/ai-hub && npm run fetch:all`	Server deployment

Recommended setup: Use npm run fetch:watch for WeChat sources (instant updates) + WebUI auto-fetch every 30-60 min for all other RSS sources.

Reinstall / Reset

If something breaks (missing commands, modules not found, etc.):

# Reinstall all dependencies from scratch
rm -rf node_modules desktop/node_modules
npm install                    # WebUI dependencies
npm run desktop:install        # Desktop widget dependencies

# Verify everything works
npm run dev                    # WebUI → http://localhost:3000
npm run desktop:dev            # Desktop widget (separate terminal)

WeChat Sources (WeWe RSS + Docker)

This section is entirely OPTIONAL. If you don't need WeChat public account content, skip this section completely. All other 60+ sources (RSS feeds, arXiv papers, etc.) work out of the box with just npm install — no Docker needed.

AI Hub uses WeWe RSS to fetch WeChat public account (微信公众号) articles. WeWe RSS is an independent open-source project that converts WeChat subscriptions into standard RSS/Atom feeds. It runs as a separate Docker container on your machine.

Why is Docker needed?

WeChat does not provide official RSS feeds. WeWe RSS acts as a bridge:

It runs a Docker container with a headless browser
It periodically scans WeChat accounts you configure
It exposes the articles as standard Atom feeds at http://localhost:4000
AI Hub then fetches these Atom feeds just like any other RSS source

Prerequisites

Docker Desktop installed and running
A WeChat account for login authentication in WeWe RSS

Setup Steps

# Step 1: Pull and start WeWe RSS container (scan every 10 minutes)
docker run -d \
  --name wewe-rss \
  -p 4000:4000 \
  -e DATABASE_TYPE=sqlite \
  -e AUTH_CODE=your_auth_code \
  -e CRON_EXPRESSION="*/10 * * * *" \
  -v $(pwd)/wewe-data:/app/data \
  cooderl/wewe-rss-sqlite:latest

# Step 2: Verify it's running
docker ps | grep wewe-rss

# Step 3: Open the dashboard
open http://localhost:4000

CRON_EXPRESSION="*/10 * * * *" means WeWe RSS scans WeChat accounts every 10 minutes. Default is every 2 hours which is too slow for timely updates.

In the WeWe RSS dashboard:

Log in with the auth code you set above
Scan the QR code with your WeChat to authorize
Add WeChat public accounts you want to follow (search by name)

How It Works

WeChat Public Accounts
        ↓ (WeWe RSS scans via authorized WeChat)
WeWe RSS (Docker, localhost:4000)
        ↓ (Atom feed: /feeds/MP_WXS_xxxxx.atom)
AI Hub engine.mjs (fetches like regular RSS)
        ↓
SQLite database → WebUI + Desktop Widget

Adding WeChat Sources to AI Hub

AI Hub comes with 10+ WeChat sources pre-configured (pointing to localhost:4000). Once WeWe RSS is running and accounts are added, they'll be fetched automatically.

To add more sources manually:

In WeWe RSS dashboard → find the feed URL (e.g., http://localhost:4000/feeds/MP_WXS_3073282833.atom)
In AI Hub → Settings → Data Sources → + Add Source:
- Name: 机器之心 (display name)
- URL: the feed URL from WeWe RSS
- Module: select target module (e.g., "AI News" or "World News")

Pre-configured WeChat Sources

Source	Category
机器之心, 新智元, 智猩猩AI, 36氪(微信), 电手, 数字生命卡兹克	AI News
人民日报, 央视军事, 九万里, 外军防务研究前沿	World News

Daily Usage

# Start WeWe RSS (run once after system reboot)
docker start wewe-rss

# Start the watcher (auto-fetches when WeWe has new content)
npm run fetch:watch

# Check WeWe RSS status
docker ps | grep wewe-rss

# Stop WeWe RSS (when not needed)
docker stop wewe-rss

# View WeWe RSS logs (if something goes wrong)
docker logs wewe-rss --tail 50

Recommended Workflow (Maximum Timeliness)

WeWe RSS (Docker)          AI Hub Watcher              Result
Every 10 min scan  →  Every 2 min check feed  →  New content in ≤12 min

# One-time setup complete, daily just run:
docker start wewe-rss && npm run fetch:watch

If WeWe RSS is not running: WeChat sources will fail silently during fetch — you'll see fewer results but no errors. All other sources continue working normally.

Desktop Widget

Run in Development Mode

# Prerequisites: Rust must be installed
curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh

# First time: install desktop dependencies
npm run desktop:install

# Start desktop widget (dev mode with hot-reload)
npm run desktop:dev

The widget will appear as a floating logo on your desktop. Click to expand the card list.

Note: The desktop widget reads from the same data/ai-hub.db as the WebUI. Make sure you've run npm run fetch:all at least once so there's data to display.

Build for Distribution

npm run desktop:build

# Output:
# .app → desktop/src-tauri/target/release/bundle/macos/AI Hub.app
# .dmg → desktop/src-tauri/target/release/bundle/dmg/AI Hub_0.1.0_aarch64.dmg

Install from DMG

Download from GitHub Releases and drag to Applications.

macOS Security Warning: If you see "AI Hub is damaged and can't be opened", run:
xattr -cr /Applications/AI\ Hub.app
This is normal for unsigned apps.

DMG Limitations: The standalone DMG includes pre-bundled data for immediate viewing and AI chat. However, fetching new data is not available without the full project directory + Node.js. For full functionality including auto-fetch, clone the repository instead.

Architecture

ai-hub/
├── src/                  # Next.js WebUI (App Router)
│   ├── app/             # Pages + API routes
│   ├── components/      # React components
│   ├── lib/             # SQLite queries, utilities
│   └── i18n/            # Bilingual translations
├── desktop/              # Tauri Desktop Widget
│   ├── src/             # React frontend
│   └── src-tauri/       # Rust backend
├── scripts/              # Data fetching engine
│   ├── engine.mjs       # Main parallel fetcher
│   ├── fetch-papers.mjs # arXiv paper fetcher
│   └── fetchers/        # RSS, scrape, API strategies
├── data/                 # SQLite database
│   └── ai-hub.db        # Shared by WebUI + Desktop
└── public/              # Static assets

Data Flow

RSS/API Sources → engine.mjs (parallel fetch + filter)
                       ↓
              SQLite (data/ai-hub.db)
                   ↙        ↘
          Next.js WebUI    Tauri Desktop Widget
              ↓                    ↓
        Browser (SSR)     Native macOS Window

Both apps read/write the same database — favorites, sources, and configuration stay in sync automatically.

Tech Stack

Layer	Technology
WebUI	Next.js 16, React 19, Tailwind CSS 4
Desktop	Tauri 2, Rust, React, Vite
Database	SQLite (better-sqlite3) with WAL mode
AI Chat	OpenAI-compatible API, SSE streaming
Fetching	rss-parser, parallel with deadline
Deploy	Vercel (WebUI), GitHub Releases (Desktop)

Documentation

Technical Documentation — Architecture deep-dive, database schema, API reference
User Guide — Feature walkthrough, configuration tips, FAQ

License

MIT

Name		Name	Last commit message	Last commit date
.github/workflows		.github/workflows
data		data
desktop		desktop
docs		docs
preview		preview
public		public
scripts		scripts
src		src
.env.example		.env.example
.gitignore		.gitignore
LICENSE		LICENSE
README.md		README.md
README.zh-CN.md		README.zh-CN.md
eslint.config.mjs		eslint.config.mjs
next.config.ts		next.config.ts
package-lock.json		package-lock.json
package.json		package.json
postcss.config.mjs		postcss.config.mjs
tsconfig.json		tsconfig.json

Latest commit

History

Folders and files

Repository files navigation

What is AI Hub?

Features

Intelligent Aggregation Engine

WebUI

Homepage

AI News Aggregation

Paper Tracker

World News

AI Products Directory

Favorites

Settings & Configuration

Desktop Widget (macOS)

Quick Start

Prerequisites

1. Clone & Install

2. One Command Start

3. (Optional) Configure LLM for AI Chat

4. Configure LLM (Optional — for AI Chat)

Available Commands

Data Fetching

WebUI

Desktop Widget

Quick Launch (Full Stack)

Auto-fetch Options

Reinstall / Reset

WeChat Sources (WeWe RSS + Docker)

Why is Docker needed?

Prerequisites

Setup Steps

How It Works

Adding WeChat Sources to AI Hub

Pre-configured WeChat Sources

Daily Usage

Recommended Workflow (Maximum Timeliness)

Desktop Widget

Run in Development Mode

Build for Distribution

Install from DMG

Architecture

Data Flow

Tech Stack

Documentation

License

About

Resources

Stars

Watchers

Forks

Releases

Packages

Contributors

Languages