getmaxun/maxun

Maxun – Turn Any Website Into a Structured API

What it is – Maxun is an open‑source web‑data platform that lets you scrape, crawl, monitor, and extract structured information from any website. It combines traditional web‑scraping with AI‑assisted extraction, letting you describe the data you want in natural language or record browser actions to build reusable “robots”.

Key capabilities

• Extract – Turn page content into structured records. Two modes:
  • Recorder: record clicks/scrolls in a browser and replay them as an extraction robot.
  • AI: write a plain‑English prompt (e.g., “list all product names and prices”) and let an LLM parse the page.
• Scrape – Save full pages as Markdown, HTML, or screenshots.
• Crawl – Walk an entire site, applying extraction robots to every page.
• Search – Query the web from within Maxun and filter results by time.
• Monitoring – Watch pages for changes and receive alerts.
• Document parsing – Extract tables, forms, or other structured data from PDFs, DOCX, etc.

How you interact with it

• Dashboard – A hosted UI (app.maxun.dev) for building and running robots without code.
• API – REST endpoints for programmatic access to scraping and extraction.
• SDK – Language‑specific libraries (e.g., a Node.js SDK) to embed Maxun in your own applications.
• CLI – Command‑line tools to create robots, trigger runs, and fetch results.
• MCP (Model Context Protocol) – A hook that lets external AI agents call Maxun as a data source.

Getting started

1. Try the hosted service – Sign up at https://app.maxun.dev and start building robots instantly.
2. Run locally – Follow the Docker‑Compose or plain‑local installation guides in the docs. The repo includes all code needed to self‑host.
3. Use the SDK/CLI – Install the Node SDK (npm i @maxun/sdk) or the CLI to script jobs from your own codebase.

Typical use cases

• Building a product‑price API by extracting listings from e‑commerce sites.
• Monitoring competitor webpages for price or content changes.
• Collecting research data from public portals that only provide HTML pages.
• Feeding up‑to‑date web content into LLM‑powered agents via the MCP integration.

License – The project is released under the AGPL‑v3, encouraging open collaboration while requiring that any network‑deployed modifications also be shared.

Community & support – The repo lists sponsors (proxy providers) and offers a Discord, tutorial videos, and documentation site. Contributions are welcomed via pull requests, and the project encourages commercial users to sponsor or give back.