xtomd API for AI Agents: Free X to Markdown Endpoint

Published March 2026 · 8 min read

X (formerly Twitter) content is everywhere, but reading it programmatically is broken. When your AI agent tries to fetch a page from x.com, it gets a JavaScript shell with no content. The server returns static HTML that requires rendering to display actual posts, threads, and articles.

That is where the xtomd API comes in. It converts any X post, thread, or article into clean Markdown that any agent can read. No API key, no authentication, no JavaScript rendering required.

The Problem Every Agent Hits

Every AI agent framework runs into the same wall when working with X content:

• Claude Code cannot read x.com URLs directly with web_fetch
• CrewAI agents get blank pages when trying to scrape threads
• LangChain tools fail at X content extraction
• AutoGPT and OpenClaw cannot parse tweets into usable text

The root cause is that X requires JavaScript execution to render content. Your serverless function, CLI tool, or agent runs server-side and gets an empty page. You need a middleman that loads X content, extracts the actual post or article text, and serves it in a format your agent can use.

That is exactly what the xtomd API does.

The API Endpoints

xtomd provides two endpoints. Pick the one that fits your workflow.

POST /api/markdown — Markdown Response (Recommended for Agents)

Endpoint: https://xtomd.com/api/markdown

Request body:

{
  "url": "https://x.com/username/status/1234567890"
}

JSON response (default):

{
  "markdown": "# Post Title\n\nPost content here...",
  "url": "https://x.com/username/status/1234567890",
  "author": {
    "name": "Username",
    "handle": "@username"
  }
}

Add the header Accept: text/markdown to get plain Markdown text instead of JSON. This is useful for piping output directly into another tool.

Use /api/markdown when feeding content directly into an LLM. It strips away noise and gives you clean, readable text.

POST /api/fetch — Raw JSON Response

Endpoint: https://xtomd.com/api/fetch

This returns the full structured data including author metadata, engagement metrics (likes, retweets, views, bookmarks), media attachments, quoted tweets, and Draft.js block content for articles. Use it when you need structured data to process rather than just the text.

Quick Test

Copy this and run it in your terminal to test the API:

curl -X POST https://xtomd.com/api/markdown \
  -H "Content-Type: application/json" \
  -d '{"url": "https://x.com/elonmusk/status/1234567890"}'

Replace the URL with any public X post. You will get back structured JSON with the Markdown content.

Integration Examples

Claude Code (One-Liner)

Claude Code can fetch X content using curl inside a bash tool call:

curl -s -X POST https://xtomd.com/api/markdown \
  -H "Content-Type: application/json" \
  -H "Accept: text/markdown" \
  -d '{"url":"https://x.com/username/status/1234567890"}'

Or use the MCP server for native integration. Install globally and add it to your Claude Code config:

npm install -g @xtomd/mcp-server

Then in your Claude Code MCP config:

{
  "mcpServers": {
    "xtomd": {
      "command": "xtomd-mcp"
    }
  }
}

Python (Works with CrewAI, LangChain, AutoGPT)

import requests

def fetch_x_content(url):
    response = requests.post(
        "https://xtomd.com/api/markdown",
        json={"url": url}
    )
    return response.json()["markdown"]

# Use it
content = fetch_x_content("https://x.com/username/status/1234567890")
print(content)

For CrewAI, wrap this in a tool decorator:

from crewai.tools import tool

@tool("Fetch X Content")
def fetch_x_thread(url: str) -> str:
    """Fetches an X post or thread and returns it as Markdown."""
    response = requests.post(
        "https://xtomd.com/api/markdown",
        json={"url": url}
    )
    return response.json()["markdown"]

Node.js

async function fetchXContent(url) {
  const response = await fetch("https://xtomd.com/api/markdown", {
    method: "POST",
    headers: {"Content-Type": "application/json"},
    body: JSON.stringify({url})
  });
  const data = await response.json();
  return data.markdown;
}

const content = await fetchXContent("https://x.com/username/status/1234567890");
console.log(content);

OpenClaw

When the bird CLI in OpenClaw fails to fetch X content (which happens frequently because X blocks server-side requests), use xtomd as a fallback. An OpenClaw skill is available that automatically detects X URLs and routes them through the xtomd API.

CLI Tool (npx)

For quick one-off conversions from your terminal:

npx @xtomd/cli https://x.com/username/status/1234567890

This prints clean Markdown to stdout. Add --json for JSON output. Pipe it anywhere.

API Response Schema

/api/markdown Response

/api/fetch Response

Rate Limits and Fair Use

• No API key required — the API is completely open
• No rate limiting currently — but be respectful of the service
• Results cached for 5 minutes — repeated requests for the same URL return instantly
• Self-hosting available — if you need high-volume access, xtomd is open source and can be deployed on your own Cloudflare Workers account

The endpoint is designed for production use. It handles thousands of requests daily. If you are building something that needs massive scale, check the GitHub repo and self-host.