scrapi — Claude Code MCP server: URLs into Markdown
Hands your agent a web page as clean Markdown or text, with headings, lists and links kept. Useful when a task starts from a link rather than a file.
claude mcp add --transport http scrapi https://scrapi.ai/mcpWhat scrapi does
scrapi is an MCP server that takes a URL and returns page content an LLM can read straight away. Metadata comes along too: title, author, date, images. JavaScript-heavy pages get rendered, so single-page apps are not a dead end.
Two transports are available. The stdio mode runs through npx and needs Node.js 20+; the HTTP mode points at https://scrapi.ai/mcp and asks for nothing installed on the client. Your key goes in SCRAPI_API_KEY or in the --api-key argument. Rate-limited responses come back with 429 and the server retries them on its own, following Retry-After.
You need a Scrapi MCP account, which is separate from the main Scrapi account, and a key starting with hsmcp_. Grab the key from the dashboard first — the server will not start without it, so editing your client config before that is wasted work.
What it helps with
- read a linked article and summarise it without ad clutter
- pull the title, author and publish date off a page
- extract text from a site that renders in JavaScript
- get a page's headings and links as a plain list
- feed documentation to an agent by URL instead of pasting it
Example prompts for the agent
- Open https://scrapi.ai/mcp and list the endpoints it exposes
- Convert this page to Markdown and save it as
notes.md - Pull the title, author and date from this link
How to install scrapi
1Platform CLI: remote server
claude mcp add --transport http scrapi https://scrapi.ai/mcp2Platform CLI
claude mcp add --env SCRAPI_API_KEY=<SCRAPI_API_KEY> --transport stdio scrapi -- npx -y @scrapi.ai/mcp-serverReplace <…> with your own values
1Platform CLI: remote server
codex mcp add scrapi --url https://scrapi.ai/mcp2Platform CLI
codex mcp add scrapi --env SCRAPI_API_KEY=<SCRAPI_API_KEY> -- npx -y @scrapi.ai/mcp-serverReplace <…> with your own values
Config file
{
"mcpServers": {
"scrapi": {
"url": "https://scrapi.ai/mcp"
}
}
}{
"mcpServers": {
"scrapi": {
"command": "npx",
"args": [
"-y",
"@scrapi.ai/mcp-server"
],
"env": {
"SCRAPI_API_KEY": "<SCRAPI_API_KEY>"
}
}
}
}Replace <…> with your own values
Config file
{
"servers": {
"scrapi": {
"type": "http",
"url": "https://scrapi.ai/mcp"
}
}
}{
"servers": {
"scrapi": {
"type": "stdio",
"command": "npx",
"args": [
"-y",
"@scrapi.ai/mcp-server"
],
"env": {
"SCRAPI_API_KEY": "<SCRAPI_API_KEY>"
}
}
}
}Replace <…> with your own values
1Platform CLI: remote server
gemini mcp add -t http scrapi https://scrapi.ai/mcp2Platform CLI
gemini mcp add -e SCRAPI_API_KEY=<SCRAPI_API_KEY> scrapi npx -y @scrapi.ai/mcp-serverReplace <…> with your own values
Step by step, with commands for each platform: How to add an MCP server to Claude Code, Codex, Cursor, VS Code, Gemini CLI and more
Questions about scrapi
How do I install scrapi in Claude Code?
Over HTTP: claude mcp add --transport http scrapi https://scrapi.ai/mcp. Or over stdio: claude mcp add --env SCRAPI_API_KEY=<SCRAPI_API_KEY> --transport stdio scrapi -- npx -y @scrapi.ai/mcp-server.
What does scrapi do?
It fetches a page by URL and returns the content as Markdown or plain text, along with metadata. Your agent reads that directly instead of you copying text by hand.
Does scrapi need Node.js?
Only for the stdio mode through npx, which requires Node.js 20+. Connecting over HTTP to https://scrapi.ai/mcp needs no Node.js on the client.
