{
  "markdown": "# ScrapeUnblocker remote MCP server\n\nA hosted (HTTP) [Model Context Protocol](https://modelcontextprotocol.io) server for\n[ScrapeUnblocker](https://scrapeunblocker.com?utm_source=mcp&utm_medium=integration&utm_campaign=mcp-remote), deployed as a Vercel serverless\nfunction. It lets **claude.ai** (web and mobile), Claude Desktop, Claude Code, and any\nother MCP client fetch any web page's HTML - or AI-parsed JSON, or Google results -\nthrough ScrapeUnblocker's anti-bot API, billed to your own account.\n\n> Prefer a local install with no hosting? Use the stdio package instead:\n> [`scrapeunblocker-mcp`](https://www.npmjs.com/package/scrapeunblocker-mcp).\n\n## Endpoint\n\n```\nhttps://mcp.scrapeunblocker.com/mcp\n```\n\nYou need a ScrapeUnblocker account - [create a free one](https://app.scrapeunblocker.com?utm_source=mcp&utm_medium=integration&utm_campaign=mcp-remote).\nThere are two ways to authenticate, and **you only need one of them**.\n\n**A. Sign in with OAuth (recommended)** - nothing to copy or paste. The server is an\nOAuth 2.1 Resource Server backed by Auth0, so the client runs a standard PKCE flow, and\nthe server resolves your ScrapeUnblocker key server-side from the account you signed in\nwith (the token is never passed through to the backend, per RFC 8707).\n\nIf your client asks for an OAuth client ID, use the public one - it is a PKCE client,\nso it carries no secret and is safe to share:\n\n```\n5BM5Wk2dE4ABkDITmuKfemPLnn3QQ8jd\n```\n\nLeave the client secret field empty.\n\n**B. Bring your own API key** - handy for scripts and clients without an OAuth flow.\nSupply the key any of three ways:\n\n1. `?key=YOUR_KEY` (or `?token=YOUR_KEY`) in the URL.\n2. `Authorization: Bearer YOUR_KEY` header (a non-JWT value).\n3. `x-scrapeunblocker-key: YOUR_KEY` header.\n\nListing the tools (`initialize`, `tools/list`) needs no credentials at all, so MCP\ndirectories and inspectors can introspect the server; actually running a tool does.\n\n### OAuth configuration (maintainers)\n\n| Env var | Purpose |\n|---------|---------|\n| `AUTH0_ISSUER` | Auth0 issuer URL, e.g. `https://TENANT.auth0.com/` (trailing slash). Enables OAuth together with `MCP_AUDIENCE`. |\n| `MCP_AUDIENCE` | Canonical MCP URI = the Auth0 API Identifier, `https://mcp.scrapeunblocker.com/mcp`. The token `aud` must match this. |\n| `MCP_EMAIL_CLAIM` | Optional namespaced claim carrying the user's email (set by an Auth0 Post-Login Action), e.g. `https://scrapeunblocker.com/email`. Falls back to the standard `email` claim. |\n| `MCP_RESOURCE_METADATA_URL` | Optional override for the RFC 9728 metadata URL advertised in `WWW-Authenticate`. |\n| `UTILS_API_BASE_URL` | utils-api base for the `email_to_key` lookup, e.g. `https://utils-api.scrapeunblocker.com`. |\n| `AWS_ACCESS_KEY_ID` / `AWS_SECRET_ACCESS_KEY` / `AWS_REGION` | IAM creds (SigV4, `execute-api`) for calling utils-api. |\n\nDiscovery endpoints served (via `vercel.json` rewrites):\n`/.well-known/oauth-protected-resource` and `/.well-known/oauth-protected-resource/mcp`.\n\n## Add it to claude.ai\n\n1. Settings → Connectors → **Add custom connector**.\n2. **URL:** `https://mcp.scrapeunblocker.com/mcp`\n3. Under **Advanced settings**, set **OAuth Client ID** to `5BM5Wk2dE4ABkDITmuKfemPLnn3QQ8jd`\n   and leave **OAuth Client Secret** empty.\n4. **Add**, then sign in and **Authorize**. The tools appear across claude.ai web,\n   mobile, and Claude Desktop.\n\nOr skip OAuth entirely and paste a personalised URL instead:\n`https://mcp.scrapeunblocker.com/mcp?key=YOUR_API_KEY`\n\n## Add it to Claude Code\n\n```bash\nclaude mcp add --transport http scrapeunblocker \"https://mcp.scrapeunblocker.com/mcp?key=YOUR_API_KEY\"\n```\n\n## Tools\n\n| Tool | What it does |\n|------|--------------|\n| `fetch_html` | Fetch the fully rendered HTML of a URL. Optionally pass `steps` to interact with the page (search, click, type, paginate) in a real browser before capture. |\n| `fetch_parsed` | Fetch a page and return AI-parsed structured JSON. |\n| `google_search` | Run a Google search and return organic results as JSON. |\n| `list_elements` | Return a page's interactive elements (buttons, inputs, selects, links, forms) with ready-to-use selectors, as JSON, to build `steps`. |\n\n### Interacting with a page (`steps`)\n\n`fetch_html` accepts an optional `steps` array - an ordered list of browser\nactions run in a real browser **after** the page loads, before the HTML is\ncaptured. Supported actions: `wait_for`, `wait_for_text`, `wait`, `click`,\n`type` (human-like keystrokes), `select`, `press_key`, `scroll`. A typical flow\nis: call `list_elements` to discover selectors, then call `fetch_html` with\n`steps` to drive the page.\n\nSteps are **non-idempotent** - the request runs once and is not retried. If a\nstep fails (bad selector, element never appeared), the tool returns an HTTP 422\nresult naming the offending step (`step_index`, `action`, `reason`, `selector`)\nplus the page HTML at the point of failure, so you can correct the step.\n\n## Deploy (maintainers)\n\nThis is a standard Vercel project - no build step, the function lives in `api/mcp.ts`.\n\n```bash\nvercel            # preview\nvercel --prod     # production\n```\n\nThen point the `mcp.scrapeunblocker.com` domain at the Vercel project. Scraping calls\ncan be slow, so the function's `maxDuration` is set to 60s (raise it on a paid Vercel\nplan if you hit timeouts on heavy pages).\n\n## License\n\nMIT\n",
  "bytes": 5350,
  "sha": "b62f40d3e37d2d75b0da46b4453f92ef630a50094ae913a1236b1e3dcdd18040",
  "repo_slug": "scrapeunblocker/scrapeunblocker-mcp-remote",
  "fonte": "repo",
  "truncated": false,
  "api": "https://agentalog.com/api/listings/mcp_com_scrapeunblocker_scrapeunblocker_mcp__cf7a9015/readme"
}