link-archiver

name	link-archiver
description	Telegram Link Archiver - Automatically summarize, tag, and save links shared via Telegram.
homepage	https://github.com/fanani/radit-link-archiver
metadata	{"emoji":"🔗","requires":{"bins":["python3","ollama"],"python_packages":["requests"]},"install":[{"id":"setup","kind":"script","command":"python3 automation/link-archiver.py --init","label":"Initialize link archiver config"}]}

🔗 Link Archiver

Auto-archive links from Telegram with AI summaries and smart tagging.

Forward any link to the Telegram bot and it will:

1. ✅ Extract and fetch the page content
2. 📝 Generate AI summary (local Ollama or cheap API)
3. 🏷️ Auto-tag by category (tech, business, engineering, etc.)
4. 💾 Save to structured Markdown storage
5. 📱 Confirm back with summary + tags

🚀 Quick Start

# Initialize config
python3 automation/link-archiver.py --init

# Test the archiver
python3 automation/link-archiver.py --test

# Process a URL directly
python3 automation/link-archiver.py "https://example.com/article"

🎯 Features

Automatic URL Extraction

• Extracts URLs from any Telegram message
• Handles multiple URLs in one message
• Cleans and validates URLs

AI-Powered Summarization

• Primary: Local Ollama (phi3:mini) - FREE, private
• Fallback: OpenRouter API (cheap models)
• Generates 2-3 sentence summaries

Smart Auto-Tagging

Automatically tags content based on:

• Keyword analysis
• Domain detection
• Content category matching

Available Tags:

• tech, ai-ml, programming, engineering
• business, startup, finance, marketing
• science, health, research
• tutorial, news, opinion
• design, productivity, entertainment

Structured Storage

• Markdown format for easy reading
• Organized by date
• Metadata included (URL, tags, timestamp)
• Stored in memory/link-archiver/articles/

Telegram Integration

• Instant confirmation messages
• Formatted with emojis
• Summary preview
• Tag display

📋 Configuration

Edit config/link-archiver.json:

{
  "telegram": {
    "bot_token": "YOUR_BOT_TOKEN",
    "chat_id": "YOUR_CHAT_ID"
  },
  "ollama": {
    "enabled": true,
    "model": "phi3:mini",
    "host": "http://localhost:11434"
  },
  "openrouter": {
    "enabled": false,
    "api_key_env": "OPENROUTER_API_KEY",
    "model": "google/gemini-3-flash-preview:free"
  },
  "storage": {
    "type": "local",
    "path": "memory/link-archiver/articles"
  },
  "tags": {
    "categories": ["tech", "business", "engineering", ...]
  }
}

🔧 Usage

Via Telegram Bot

1. Forward any message with a link to @RaditClaw_bot
2. Bot processes automatically
3. Receive confirmation with summary + tags

Command Line

# Single URL
python3 automation/link-archiver.py "https://news.ycombinator.com/item?id=123"

# From message text (with context)
python3 automation/link-archiver.py "Check this out: https://example.com/article cool stuff"

# Test mode
python3 automation/link-archiver.py --test

# Re-initialize config
python3 automation/link-archiver.py --init

Python Import

from automation.link_archiver import process_message, load_config

config = load_config()
result = process_message("https://example.com", config)
print(result['summary'])
print(result['tags'])

📁 Output Format

Saved articles as Markdown:

# Article Title

**URL:** [link](url)  
**Archived:** 2026-03-13T10:30:00  
**Tags:** #tech #ai-ml #programming

## Summary

AI-generated summary of the article content...

## Original Content

First 500 chars of original article...

---
*Auto-archived by Radit Link Archiver*

🔄 Integration with Radit Bot

The link archiver integrates with the existing Radit Telegram bot:

1. Message Handler: skills/link-archiver/scripts/telegram-handler.py
2. Webhook/Pooling: Works with existing bot infrastructure
3. Command: /archive to manually trigger archiving

🛠️ Requirements

• Python 3.8+
• Ollama (for local AI): curl -fsSL https://ollama.com/install.sh | sh
• At least one Ollama model: ollama pull phi3:mini
• requests library: pip3 install requests
• summarize CLI (optional, for better extraction): brew install steipete/tap/summarize

📊 Storage Structure

memory/link-archiver/
├── articles/
│   ├── 2026-03-13-how-to-build-ai-apps.md
│   ├── 2026-03-13-startup-funding-news.md
│   └── ...
└── index.json (article registry)

🔍 How It Works

Telegram Message
       ↓
Extract URLs
       ↓
Fetch Page Content (summarize CLI → fallback to requests)
       ↓
Generate Summary (Ollama → OpenRouter fallback)
       ↓
Auto-Tag (keyword + domain analysis)
       ↓
Save to Markdown
       ↓
Send Telegram Confirmation

🐛 Troubleshooting

No summary generated?

• Check Ollama: ollama list
• Test Ollama: ollama run phi3:mini "Summarize this"
• Check OpenRouter API key if using fallback

Page not fetching?

• Some sites block bots (use summarize CLI with Firecrawl)
• Check URL is accessible: curl -I <url>

Telegram not sending?

• Verify bot token: Test with curl
• Check chat_id is correct
• Ensure bot has permission to send messages

📝 TODO

• Supabase storage backend
• GitHub storage backend
• Web UI for browsing archives
• Search functionality
• Export to PDF/Notion
• Duplicate detection
• Readwise integration

🙏 Credits

• Summarization: Ollama (local) / OpenRouter (API)
• Content extraction: summarize.sh CLI
• Part of Radit Assistant ecosystem