How to get X (Twitter) data into your AI agent through MCP.

Agents got good this year. They can reason and they can call tools. If you run Claude Code or Cursor daily, you have probably asked one to help with your X account already. Then you hit the same wall every builder hits. The agent has no idea what is happening on your account. It cannot see which posts pulled 40k impressions and which died at 200. It does not know your voice. And it cannot ship a post unless you copy and paste it.

This post is the practical fix. It covers what MCP is and why it beats the DIY routes for X data. It shows exactly what your agent can read and do through the ClimbX MCP server, plus the one-command setup for each major client. It ends with the customer who built the first MCP server on our API before we shipped ours, and what that says about where product access is going.

What is MCP, and why does it matter for X data?

Model Context Protocol is an open standard for connecting AI agents to tools and data. Think of it as USB for agents. Without it, every product writes a custom plugin for every agent. With it, a product runs one MCP server. Then every client that speaks MCP can find and call its tools with no extra work. That includes Claude Code, Claude Desktop, Cursor, OpenAI's Agents SDK and a fast-growing list of others.

This matters for X data because X data is annoying to get. The three DIY routes all have a catch:

• Scraping breaks every few weeks and breaks the X terms of service. It also can never show you your own analytics (impressions are not public).
• The official X API is reliable but priced per resource. Reading a single post costs $0.005. A post with a URL costs $0.200 to publish, 13x the standard $0.015 (X API pricing). Setting up OAuth 2.0 with token refresh in an agent eats a weekend before the first useful call.
• Raw data anyway. Even once it is set up, the API gives you posts and counts, not answers. "Which of my formats is trending up this month" is a pipeline you would still have to build. You need classification, baselines, medians per format and trend windows.

An MCP server on a product that already solved those three problems skips all of that. The agent asks the kind of question you would ask ("what worked for me last month?"). It gets back data it can reason over.

What X data can your agent actually get?

The ClimbX MCP server has 16 tools in four groups. Together they cover the loop an agent needs to be really useful on X. See what works, draft from it, ship it, measure it.

Analytics, voice, and learnings. Headline KPIs over any window up to 90 days. Performance per format and per niche, with medians and trends (so "hot takes are up, threads are flat" is one tool call). Your voice profile with cadence targets. The do-more / do-less rules ClimbX built from your own posts, plus how those rules changed over time. An agent needs this data before it drafts anything. Scraping can never give it.

Inspiration outliers. These are the posts breaking out right now for the creators you track. Each one has a multiplier over that author's own baseline. A 3.4x post is a real outlier whether the account is at 2k or 200k followers. There is also a discovery feed of outliers from across the network. You can filter it by multiplier, impressions, format and recency. Outliers beat raw timelines as agent input because the filtering is already done. Every item is a proven breakout, not noise.

Publishing and scheduling. Publish a post now or schedule one for later (ISO datetime, timezone-aware). List what is queued, move it or cancel it. The agent does not have to stay awake to post at 8am. The scheduling pipeline handles on-the-minute delivery and retries.

Reply drafting. One tool drafts a short reply in your voice to a post you might reply to. It returns a suggestion for you to edit and send yourself. It never posts a reply. AI reply spam is the most algorithm-penalized behavior on X, and we are not building it.

How do you connect Claude Code, Claude Desktop, or Cursor?

Create an API key in ClimbX settings. Settings -> MCP shows these snippets with your key already filled in. For Claude Code, the whole setup is one command:

claude mcp add --transport http climbx https://climbx.so/mcp \
  --header "Authorization: Bearer climbx_sk_your_key"

Cursor and Claude Desktop take a JSON config block instead. Same URL, same header:

{
  "mcpServers": {
    "climbx": {
      "url": "https://climbx.so/mcp",
      "headers": { "Authorization": "Bearer climbx_sk_your_key" }
    }
  }
}

Any other MCP client that supports remote HTTP servers works the same way. There is nothing to install and nothing to keep running. The server is hosted, and the key is all you need to connect.

What can an agent actually do with it?

Here are three workflows we see people use. All of them work the minute the server is connected:

The morning briefing. "Check my analytics for the last 30 days, pull the top outliers from the creators I track, and draft three posts in my voice about whichever topic is breaking out." The agent chains get_analytics, get_following_outliers, and get_voice_profile. It hands you drafts based on what is working this week for accounts your size. No generic AI slop.

The scheduler. "Here are my five post ideas for the week. Check my posting schedule and queue them at my best times." The agent reads the cadence targets from your voice profile. It schedules each post with schedule_post. You review the queue in the ClimbX calendar like any other scheduled post.

This workflow is X-only on purpose. If you need one agent to schedule and publish across several social networks, AgentReacher is a multi-platform option for managing those channels from one place.

The performance review. "Compare my format performance this month vs last month, and tell me what the learnings history says changed." You get trends per format and niche, plus how the do-more / do-less rules changed. All in one conversation, not an evening digging through spreadsheets.

What stops an agent from wrecking your account?

The real risk of giving an agent publish access is a runaway loop. It could ship nonsense to the account you spent a year building. So the limits are set on purpose and enforced on the server, where a broken agent cannot get around them:

• No link posts. The API rejects URLs. One reason is that a link cuts a post's reach by 50 to 90%. The other is that URL posts cost 13x to publish. The agent posts the thought. You drop the link in a reply.
• Five posts per day through the API, shared between publish and schedule. That is enough for a real cadence. It is also low enough that a loop cannot flood your feed or trip X's spam filter.
• Replies are suggestions. The reply tool returns a draft and stops. Sending it is a human action on X, every time.
• Keys you can kill. Each key works for one account. It is shown once and stored only as a hash. You can revoke it instantly, and every client using it stops on the next call.

Here is why you can trust these limits. The MCP server has zero business logic of its own. Every tool calls the exact same API route handlers as a direct REST request, with the same bearer key. There is no second copy of the logic to drift out of sync. So everything in the API docs describes the MCP tools too.

Can you build your own MCP server on top of a REST API?

Yes, and one of our customers proved it before we did. Weeks after we shipped the public API, a paying customer built a community MCP server wrapping it and put it on GitHub. There is no clearer sign that people want an API. It is what moved the official server up the roadmap.

If you run a product with a REST API and are thinking about the same move, steal what worked for us. Make every MCP tool a thin wrapper that calls your existing API handlers with the caller's own credentials. No new business logic, no separate permission model, no drift. Ours took days, not weeks. That is because the API already enforced everything that mattered. The MCP layer only translates.

Frequently asked questions

One short version

Your agent thinks. ClimbX knows X. One command connects them. Your agent gets the same data you see in the app: analytics, voice, learnings, outliers and publishing. The guardrails that keep the account yours sit on the server, where an agent cannot argue with them. Grab a key in settings and run the first briefing today.