LinkIntel

npx add-mcp 'https://www.getlinkintel.com/api/mcp'

LinkIntel

Your agent already writes your posts. Now it remembers which ones worked.

LinkIntel is an MCP server that gives your AI assistant evidence for its next social post. Once you connect X, it measures your own recent original posts daily and stores the snapshots. You can also hand it public X posts to learn from, stored with their public performance and source links. Your assistant (Claude, or any MCP-compatible client) reads both back as a cited content brief. It costs $39 USD/month and never publishes anything.

Who it is for

People who already draft posts with an AI assistant and want those drafts grounded in posts that actually worked, instead of the assistant's general idea of what performs on X. Starting with X; LinkedIn is coming to the broader workflow.

How it works

1. Connect your assistant. Add https://www.getlinkintel.com/api/mcp as a remote MCP server. OAuth 2.1 opens the browser for Google sign-in; Google does not charge you.
2. Activate your account. $39 USD/month, billed monthly in advance, no free trial. The amount is shown and requires confirmation before the assistant can use LinkIntel.
3. Add evidence. Connect X from your account page, or from the link your assistant offers, so your own posts are measured. Optionally give it public X post URLs, or an account handle to pull recent posts from; it confirms the list with you before storing anything.
4. Ask for a brief. The assistant reads the stored snapshots and references and tells you what the evidence supports, what it does not, and what is worth testing in your next draft.

Install

• Claude Code: claude mcp add -s user --transport http linkintel https://www.getlinkintel.com/api/mcp. Then start a new session (/exit, then claude --continue) and the browser opens for sign-in and payment.
• Claude.ai or Claude Desktop: Settings, then Connectors, then Add custom connector, and paste https://www.getlinkintel.com/api/mcp. The first LinkIntel request shows a Connect button.
• Any other MCP client with OAuth 2.1: add https://www.getlinkintel.com/api/mcp as a remote HTTP server.
• A client without OAuth support: create an account at https://www.getlinkintel.com/connect, then copy the private bearer-token handoff shown on https://www.getlinkintel.com/setup into the assistant.

Current X boundary

• LinkIntel reads public data for X references the customer explicitly provides or agrees to. If the customer connects X, it also reads their own recent original posts and metrics once a day and stores snapshots; disconnecting deletes them.
• LinkIntel never asks for an X password, searches or discovers posts on its own, or publishes on the customer's behalf.
• An import is limited to 100 URLs from the browser form and 25 per call from an assistant. Cached references are reused where possible.
• Reference lookups are included in the subscription and are never billed to the customer per post. Fair use: up to 100 post lookups a day and 1,000 a month per account.
• A brief is evidence, not a prediction: observed public reach does not predict your reach or prove that a writing choice caused an outcome.

LinkedIn

LinkedIn analytics is an optional source on the account page for existing customers. It is not required for the X workflow.

---

About LinkIntel

LinkIntel is an MCP server that gives an AI assistant evidence for its next social post. Once the customer connects X it reads their recent original posts and metrics daily and stores snapshots. It also stores public X posts the customer chooses, with source links, and serves X's open-source ranking weights, read daily. The assistant gets all of it as a cited content brief. LinkIntel never publishes and never searches X on its own; an X lookup is only made for posts the signed-in customer supplies and confirms, and lookups are included in the subscription, not billed per post.

Buy: https://www.getlinkintel.com/connect · Pricing: https://www.getlinkintel.com/pricing.md · Docs: https://www.getlinkintel.com/docs/mcp · Full index: https://www.getlinkintel.com/llms.txt