Back to blog

MCP

Connect an AI Agent to looktme with MCP

Learn how to connect an AI agent to looktme with MCP. Review the current developer documentation for setup requirements and supported actions.

By looktme · Published · 1 min read

Connect a compatible AI client to https://api.looktme.com/mcp using the setup steps on looktme Developers. Authenticate with Authorization: Bearer ltk_… and inspect tools before running write actions. The MCP endpoint is Streamable HTTP; OAuth sign-in is not supported.

Setup checklist

  1. Create an API key with the required scope in Settings → API keys.
  2. Configure the server endpoint and key in your client using /developers.
  3. Store credentials securely, not in prompts, source code, or shared chat.

Key safety: Keep API keys server-side. Never place real credentials in browser code, prompts, or public repositories.

  1. Start with list_connections or list_posts (read scope) and inspect agentAccess.
  2. create_draft prepares content. Publish and schedule LinkedIn and X posts from the looktme app; agent publishing is not available for them yet.

Available tools are list_connections, list_posts, get_post_status, create_draft, and schedule_post. Agent posts are drafts by default. schedule_post cannot publish or schedule to LinkedIn or X yet. MCP does not bypass platform permissions or review.

Need help? Visit looktme Help.

FAQ

Does connecting MCP automatically publish posts?

No. The effect depends on the tool you invoke; review its description and confirm the action.

Where are the current connection details?

See looktme Developers for the live endpoint and client-specific configuration.

Related guides