maxun

Maxun is an open-source, self-hostable no-code web data platform designed to turn websites into structured, reusable data sources. Users can create automated robots without writing code for extraction, page scraping, site crawling, web search, and document parsing. The Extract feature can record browsing actions and convert them into a reusable robot, or use AI Mode where users describe the desired data in natural language. It supports pagination, scrolling, scheduled runs, authenticated pages, and automatic recovery when website layouts change. Extracted data can be exposed through RESTful endpoints or exported to Google Sheets and Airtable. Scrape robots convert full webpages into clean Markdown or HTML and can capture screenshots, producing output intended for AI workflows, agents, and document processing. Crawl robots traverse websites and extract content from discovered pages while controlling crawl scope. Search robots perform automated web searches, including time-based filters, and can collect or scrape the resulting pages. Maxun also handles document and image extraction. It can parse PDF, DOCX, XLSX, and CSV files; use OCR for scanned PDFs, JPGs, and PNGs; convert files into Markdown, HTML, links, or summaries; and extract structured fields with AI assistance. Example uses include receipts, invoices, paper forms, screenshots of tables, and scanned documents. For developers, Maxun provides an SDK and CLI for creating robots, triggering runs, managing schedules, and retrieving data. It also includes Model Context Protocol support and integrations for connecting extracted data with other tools. The project is AGPLv3 licensed. It offers a hosted application at app.maxun.dev and can be installed locally or self-hosted using Docker Compose or a non-Docker setup, with documentation for environment variables and upgrades.