MCP

Use ContentExtractAPI inside Claude, Cursor, and any MCP-compatible agent.

Three tools — convert_webpage, convert_html, and bulk_convert — installable in one command. Anonymous usage works without a key; create a free key for 100 credits.

Quick install

# Test the server directly
CONTENTEXTRACT_API_KEY=ce_your_key npx contentextract-mcp

Claude Desktop

Edit ~/.claude/claude_desktop_config.json (macOS: ~/Library/Application Support/Claude/claude_desktop_config.json):

{
  "mcpServers": {
    "contentextractapi": {
      "command": "npx",
      "args": ["-y", "contentextract-mcp"],
      "env": {
        "CONTENTEXTRACT_API_KEY": "ce_your_key"
      }
    }
  }
}

Restart Claude Desktop. The three tools appear in the tool picker automatically.

Cursor

Edit ~/.cursor/mcp.json (or open Cursor → Settings → MCP):

{
  "mcpServers": {
    "contentextractapi": {
      "command": "npx",
      "args": ["-y", "contentextract-mcp"],
      "env": {
        "CONTENTEXTRACT_API_KEY": "ce_your_key"
      }
    }
  }
}

Available tools

// convert_webpage
// Fetch a public URL and return clean Markdown, text, or JSON.
// Use for: reading web pages before reasoning, RAG ingestion, competitor monitoring.
{ url: "https://example.com", format: "markdown" }

// convert_html
// Convert raw HTML (already fetched) without any network request.
// Use for: processing scraped HTML, cleaning content before storing.
{ html: "<article>...</article>", format: "text", sourceUrl: "https://example.com" }

// bulk_convert
// Convert 1–20 URLs at once. Returns a job summary.
// Use for: batch ingestion, sitemap crawls, monitoring multiple pages.
{ urls: ["https://example.com", "https://docs.example.com/guide"] }

Direct HTTP (no MCP server required)

Any agent framework that supports HTTP tool calls can use the API directly without the MCP server.

# Get a free key (100 credits)
curl -X POST https://contentextract-5f53349oz-shibley-burnetts-projects.vercel.app/api/keys \
  -H "content-type: application/json" \
  -d '{"label":"my agent"}'

# Convert a URL to Markdown
curl -X POST https://contentextract-5f53349oz-shibley-burnetts-projects.vercel.app/api/convert \
  -H "authorization: ce_your_key" \
  -H "content-type: application/json" \
  -d '{"url":"https://example.com","format":"markdown"}'

# Python (requests)
import requests
r = requests.post(
    "https://contentextract-5f53349oz-shibley-burnetts-projects.vercel.app/api/convert",
    headers={"authorization": "ce_your_key"},
    json={"url": "https://example.com", "format": "markdown"}
)
markdown = r.json()["output"]

Machine-readable docs

A llms.txt file and agent-tool-manifest.json are available for automated tool discovery.