Maxun: No-Code Web Scraping Platform
/maxun
Maxun is a self-hosted, no-code platform that turns websites into structured data. If Playwright or Puppeteer scripts breaking on every layout change are wearing you down, its browser-recorder approach is worth trying — you rebuild the robot visually instead of rewriting selectors. The catch: it's infrastructure you deploy and operate, not a library you import. Skip it for one quick one-off scrape.
What is Maxun
Maxun is an open-source, self-hosted platform for web scraping, crawling, search, and AI-driven data extraction that turns websites into structured APIs. It ships four robot types — Extract, Scrape, Crawl, and Search — plus a CLI and SDK, and its automation runs on Playwright under a TypeScript codebase.
Key features of Maxun
- ✓Recorder Mode captures your clicks and turns them into a reusable extraction robot — no selectors to write by hand.
- ✓AI Mode lets you describe what you want in plain language and hands the extraction off to an LLM.
- ✓Scrape robots convert a page into clean Markdown or HTML and can capture a screenshot in the same run.
- ✓Crawl robots walk an entire site and pull content from every page in scope, not just one URL.
- ✓Search robots run automated queries and filter results by time.
- ✓Every robot can export to a generated REST endpoint, or straight into Google Sheets or Airtable.
- ✓MCP support wires robots directly into agent tooling that speaks the Model Context Protocol.
- ✓Robots auto-recover when a site's layout changes and can run on a schedule.
What you can build with Maxun
- •Building a lead list by extracting company and contact data from directory sites.
- •Feeding LLM-ready Markdown from Scrape robots into a RAG pipeline or agent workflow.
- •Market research — tracking product prices or listings across competitor sites on a schedule.
- •Content aggregation, pulling articles or listings from many pages into one dataset via Crawl.
- •Piping scraped results straight into a Google Sheet or Airtable base instead of a database.
Installing Maxun locally
The fastest way to try Maxun is the hosted version at app.maxun.dev — no install. To self-host, Maxun runs locally with or without Docker: the README points to a Docker Compose setup guide and a separate no-Docker setup guide, plus a page for configuring environment variables. There's also a self-host guide and separate upgrade instructions for both the Docker Compose and local setups. Beyond linking to these docs, the README doesn't spell out the exact commands, so the precise steps aren't clearly documented here.
Why choose Maxun
- ✓No-code Recorder Mode means non-developers can build a working scraper.
- ✓AI Mode removes the need to hand-write extraction logic for every new site.
- ✓Four distinct robot types (Extract, Scrape, Crawl, Search) cover more than simple page scraping.
- ✓Self-hosted means your scraping infrastructure and extracted data stay on your own servers.
- ✓Built-in export to Google Sheets and Airtable skips a separate integration step.
Limitations and considerations
- △AGPL-3.0 licensing means modifying and running Maxun as a network service can trigger source-disclosure obligations — check this before embedding it in a commercial product.
- △Self-hosting means you own the Playwright/browser infrastructure yourself — headless browser workloads are memory-hungry and need real server resources, not a free-tier VM.
- △The README doesn't document exact install commands, so first-time setup means following external docs rather than a copy-paste quick start.
- △AI Mode extraction depends on an LLM you configure and pay for separately — it's not a bundled free capability.
Maxun alternatives
Frequently asked questions
Maxun is open source and free to self-host, released under the AGPL-3.0 license. That license means if you modify Maxun and run it as a network service, you're generally expected to share those source changes too — worth checking before folding it into a commercial product.
Maxun's robots run on Playwright, a real browser engine, so Extract, Scrape, and Crawl robots render JavaScript and execute page scripts instead of just fetching raw HTML. That's the same mechanism Playwright-based scrapers use to handle single-page apps.
Maxun's Recorder Mode replays the exact browser actions you clicked through once to build a robot. AI Mode skips that recording step — you describe the data you want in plain language, and an LLM drives the extraction instead of a fixed sequence.
Maxun doesn't require coding for its core workflow: Recorder Mode and AI Mode both build robots through point-and-click actions or natural-language prompts. A CLI and SDK are also available for teams that want to script extraction and scheduling.
Maxun can be self-hosted without Docker — the README links to a dedicated no-Docker setup guide alongside the Docker Compose path. Both are self-hosting options; the README doesn't spell out the exact non-Docker commands itself, only the docs link.
Extract robots emulate user behavior to pull structured data from specific elements on a page. Scrape robots convert a whole page into clean Markdown or HTML plus a screenshot. Crawl robots walk an entire site across many pages. Search robots run automated queries with time-based filters.
The problem it solves
Custom scraping scripts built with Playwright or Puppeteer selectors break every time a target site changes its markup, and someone has to notice and rewrite them by hand. Maxun's answer is to record the browsing action once, lean on AI Mode or a robot's replay logic to survive layout drift, then turn that recording into a scheduled job with a generated REST endpoint instead of a one-off script sitting in a repo.
Who should try it — and who should skip
Try Maxun if you need scraped data flowing into a spreadsheet, an API, or an LLM pipeline and you'd rather record a browser flow than maintain Playwright scripts — especially if a non-developer on the team needs to build or edit robots. Skip it if you need a single quick one-off scrape and don't want to stand up and operate a self-hosted app, or if AGPL-3.0 licensing is a blocker for how you plan to ship the result.
Related repositories
Want a second opinion on maxun?
Ask an AI that can read this page — one click and you get its take on maxun.