MCP

Contrie exposes a hosted Model Context Protocol server so agents can extract structured data as a tool call instead of writing HTTP client code. Streamable HTTP at:

https://contrie.com/mcp

Tools

Tool	Input	Returns
contrie_extract	{ url, extract?, schema? }	The same JSON response as /api/v1/scrape, as a text content block. Marked `isError` when the pipeline returned an error code.
contrie_read	{ url }	The page as markdown text. No model involved, `costUsd` 0.

Authentication

Pass your key as an Authorization header on the MCP connection, same as the REST API. Without a key, only the sample URLs succeed — see Authentication for how to get one.

Never put a key in a URL
Never paste a key into an agent chat — pass it as a connection header

Claude Code

shell

claude mcp add --transport http contrie https://contrie.com/mcp \
  --header "Authorization: Bearer $CONTRIE_API_KEY"

JSON Config (Claude Desktop, other MCP clients)

JSON

{
  "mcpServers": {
    "contrie": {
      "url": "https://contrie.com/mcp",
      "headers": {
        "Authorization": "Bearer ck_live_..."
      }
    }
  }
}

Local stdio Alternative

For clients that only speak stdio, or to point at a local or preview deployment, run the server from a clone of the repo:

shell

CONTRIE_API_KEY=ck_live_... npx -y tsx packages/mcp/src/server.ts

Set CONTRIE_API_URL to target a different deployment (defaults to contrie.com). The hosted endpoint above exposes the same two tools with nothing to install — prefer it unless your client requires stdio.