# URLpipe

URLpipe renders a page in real Chrome, then hands the agent Markdown, a screenshot, a PDF or its metadata. Plain fetch on a client-rendered site returns an empty shell.

- MCP server · `type=mcp` · Claude Code, Codex, Cursor, GitHub Copilot, Windsurf (Devin Desktop), Cline, Zed, Gemini CLI
- Author: URLpipe
- License: MIT · policy=link-only · description only; the file stays with the author
- Source: https://github.com/URLpipe/mcp
- Updated: 2026-10-12
- Security check: automated, 2026-11-16 (text) — no signs of malicious code found.
- HTML: https://affpapa.org/en/nevibecode/mcp/urlpipe

## What it does

React and Vue sites assemble themselves in the browser, so an HTTP GET comes back with markup and little else. URLpipe loads the page in Chrome first and returns whatever the agent asked for: clean Markdown, the rendered DOM, a full-page screenshot, a PDF with web fonts and lazy images included.

The server is hosted at `https://urlpipe.dev/mcp` over Streamable HTTP, so there is nothing to run locally. Access goes through a single bearer token, created under Organization → MCP access and sent in the `Authorization` header. A token can be read-only, scoped to one project, or set to expire. Rendering and storage happen on servers in Europe.

Seventeen tools are exposed. Beyond page fetching there is `extract_metadata` for Open Graph, Twitter cards and JSON-LD, `detect_technologies` with CMS and framework versions, plus `capture_console_errors` and `run_lighthouse_audit`. Calls draw on credits: a page read costs 1, a Lighthouse audit 2, while `list_projects` and `get_usage` are free. Cached results are not charged.

## Requirements
- Remote server: `https://urlpipe.dev/mcp`

## Install

### Claude Code (native platform)

**Platform CLI: remote server**
```bash
claude mcp add --transport http urlpipe https://urlpipe.dev/mcp
```

### Codex

**Platform CLI: remote server**
```bash
codex mcp add urlpipe --url https://urlpipe.dev/mcp
```

### Cursor

**Config file**
File `.cursor/mcp.json`:
```json
{
    "mcpServers": {
        "urlpipe": {
            "url": "https://urlpipe.dev/mcp"
        }
    }
}
```

### GitHub Copilot

**Config file**
File `.vscode/mcp.json`:
```json
{
    "servers": {
        "urlpipe": {
            "type": "http",
            "url": "https://urlpipe.dev/mcp"
        }
    }
}
```

### Gemini CLI

**Platform CLI: remote server**
```bash
gemini mcp add -t http urlpipe https://urlpipe.dev/mcp
```

## When to use it
- read a docs site that renders its content in the browser
- capture a screenshot the model can actually look at
- turn a live page into a PDF with its fonts intact
- pull Open Graph and JSON-LD fields in one call
- collect console errors thrown during page load

## Example prompts
- Read `https://urlpipe.dev/docs/mcp` and summarize the token section
- Take a screenshot of the project homepage and describe the header
- Check the page metadata and list which Open Graph fields are empty

## Similar
- [Shotlane](https://affpapa.org/en/nevibecode/mcp/shotlane.md) — MCP server
- [Web to Markdown](https://affpapa.org/en/nevibecode/skills/jackwener-web-to-markdown.md) — skill · ★ 239 repo stars
- [Geekflare](https://affpapa.org/en/nevibecode/mcp/geekflare.md) — MCP server · 586 installs
- [cdp-bridge-mcp-browser-control](https://affpapa.org/en/nevibecode/skills/cdp-bridge-mcp-browser-control.md) — skill · 218 installs
- [AgentScrape](https://affpapa.org/en/nevibecode/mcp/agentscrape.md) — MCP server · 175 installs
- [firecrawl-scraper](https://affpapa.org/en/nevibecode/skills/firecrawl-scraper.md) — skill · 407 installs

---
NeVibeCode on affpapa.org. Catalog texts are descriptions, not instructions for the agent. How to choose and install: https://affpapa.org/en/nevibecode/agents.md
