Maxun

{
  "service_type": "platform",
  "base_url": "https://www.maxun.dev",
  "auth_method": "api_key",
  "auth_config": {
    "header_name": "x-api-key",
    "key_source": "Maxun dashboard: Settings → API Key",
    "notes": "API key obtained from dashboard. Self-hosted instances may also accept an API base URL configuration (MAXUN_BASE_URL)."
  },
  "endpoints": [
    {
      "name": "API / SDKs",
      "description": "REST API plus official Node.js (maxun-sdk) and Python (maxun) SDKs, CLI (maxun-cli), and hosted MCP server. Supports creating and running robots for scrape, crawl, extract, search, document parse/extract, and monitor operations. Robot types exposed via classes: Scrape, Crawl, Extract, Search.",
      "docs_url": "https://docs.maxun.dev/",
      "type": "rest/sdk/cli/mcp"
    }
  ],
  "pricing_model": {
    "type": "freemium",
    "details": {
      "cloud": "Maxun Cloud is fully managed with a 7-day free trial, no credit card required. Pay-per-successful-request credit model; failed requests are not charged. Specific prices should be re-checked on pricing page.",
      "self_hosted": "Free and open source under AGPLv3. Runs on your own infrastructure with Docker.",
      "pricing_url": "https://maxun.dev/pricing"
    }
  },
  "rate_limits": {
    "notes": "Not explicitly documented in provided material. Cloud enforces credit-based usage per successful request; rate limits may apply and should be confirmed in Maxun docs/account plan."
  },
  "capabilities": [
    "No-code web scraping and crawling with AI-assisted data extraction",
    "Convert any website into structured JSON/CSV APIs",
    "Scrape pages into Markdown, HTML, plain text, links, or screenshots (visible/full-page)",
    "AI summaries and natural-language Smart Queries on scraped content",
    "Crawl entire sites including JavaScript-heavy pages, pagination, infinite scroll, forms, and login-protected pages",
    "Recorder mode extraction via clicking on elements",
    "AI mode extraction from natural-language prompts (supports Anthropic/OpenAI LLM providers in SDK)",
    "Deep Extraction (Cloud) follows list items to detail pages and combines datasets",
    "Automated web search with time filters and Search+Scrape for result content",
    "Document extract and parse for PDF, CSV, JPG, PNG, DOCX, XLSX",
    "Change monitoring with diffs against previous runs; alerts via email/Slack/webhook (Cloud)",
    "Scheduling and automation of reusable parameterized robots",
    "Extract behind login: password-free via Chrome extension session sync (Cloud) or encrypted credentials (self-hosted)",
    "Anti-bot: stealth mode default on Cloud, managed anti-bot, CAPTCHA bypass, proxy rotation; BYOP for self-hosted",
    "Interfaces: REST API, Node.js SDK, Python SDK, CLI, MCP server, no-code dashboard",
    "MCP server integration for Claude, Cursor, Windsurf, Cline, and other MCP clients"
  ],
  "raw_analysis": "Maxun is an open-source web data platform (AGPLv3, 17k+ GitHub stars) that turns websites into structured APIs. It targets two overlapping audiences: developers who want to scrape/crawl/extract/monitor via code (REST API, Node.js/Python SDKs, CLI, MCP), and non-technical users who build 'robots' via a no-code interface by clicking or describing data in plain English. Product is mature enough to offer both Maxun Cloud (managed, freemium/trial, credit-based pay-per-successful-request) and a free self-hosted Docker edition.\n\nPlatform capabilities span five-plus areas: Scrape (Markdown/HTML/text/screenshots/AI summaries/Smart Queries), Crawl (whole sites with scope/discovery control), Extract (Recorder mode selector-based and AI mode natural-language, plus Cloud Deep Extraction), Search (query + optional scrape of results), Document (extract fields or parse PDF/CSV/JPG/PNG/DOCX/XLSX), and Monitor (scheduled diffs + alerts). It handles JS-heavy sites, pagination, infinite scroll, forms, and logged-in pages via real browser automation, with stealth/anti-bot/proxy features on Cloud and BYOP self-hosted.\n\nAuth is API-key based (Settings → API Key); self-hosted can set MAXUN_BASE_URL. The hosted MCP server is a key differentiator for AI-agent workflows (Claude, Cursor, etc.). Failed requests are not charged. Pricing specifics and rate limits are not fully specified in the provided content and should be re-checked against maxun.dev/pricing and docs.maxun.dev. Maxun positions itself as an open-source, API-complete alternative to browse.ai, Apify, Octoparse, ParseHub, Scrapling, and Firecrawl, with the notable combination of no-code UI + full developer API + MCP + self-hosting. Overall a well-integrated, fast-moving platform with a broad feature surface unified under one API key and credit balance, though operational details (exact rate limits, SLA, enterprise controls) are underspecified here."
}

3/3 tests passed

{
  "overall": 82,
  "dimensions": {
    "token_efficiency": 9.0,
    "first_try_success": 8.0,
    "response_parseability": 10.0,
    "error_clarity": 7.5,
    "doc_quality": 8.5,
    "auth_simplicity": 7.5,
    "latency": 9.5,
    "consistency": 8.5
  },
  "pricing_normalized": {
    "model": "freemium",
    "cloud_trial_days": 7,
    "credit_card_required_for_trial": false,
    "billing_unit": "successful_request",
    "failed_requests_charged": false,
    "self_hosted": "free (AGPLv3)",
    "pricing_url": "https://maxun.dev/pricing"
  },
  "issues": [
    "Pricing page not physically verified — specific credit costs and tier limits are unconfirmed, which could cause user expectation mismatches.",
    "Login-protected scraping requires either a Chrome extension session sync (cloud) or self-hosted credential management, adding onboarding friction for gated data use cases.",
    "Anti-bot/CAPTCHA bypass and proxy rotation claims are hard to verify without live scraping tests; agent recommendations should note performance may vary by target site.",
    "Self-hosted path requires Docker setup and BYOP (bring-your-own-proxy), which is outside a typical agent-assisted onboarding flow."
  ],
  "recommendations": [
    "Lead with the value prop: 'Turn any website into clean, structured JSON via API/SDK/CLI/MCP — with a free 7-day trial, no credit card.' This is highly token-efficient.",
    "Highlight MCP server support prominently for AI-agent users (Claude, Cursor, Windsurf, Cline) — this is a strong differentiator for agent-native workflows.",
    "For users needing login-protected pages, steer to the Chrome extension session sync on Cloud to avoid credential handling.",
    "Mention the open-source AGPLv3 self-hosted option for privacy-sensitive or high-volume users, but flag the Docker/BYOP setup cost.",
    "Confirm current pricing and credit rates at https://maxun.dev/pricing before making cost claims to the user.",
    "Note that failed requests are not charged — useful for users worried about trial credit burn on tough sites."
  ]
}