Cursor MCP: find conversations, draft posts

Cursor is an AI code editor that talks to MCP servers, which means the window where you build the product can also run the marketing for it. Connect Oetzi and Cursor's agent can pull the conversations worth answering, draft a reply in your voice, check whether AI engines name you, and publish a post through your channels. Every paid call is quoted before it spends.

Why point a code editor at a marketing server?

Because the founder in Cursor is usually the same person who owes the market a post, and context switching is the tax. Cursor's agent already has your repo, your changelog, and the feature you just merged, and Cursor adds MCP servers from its own settings. The Oetzi MCP server supplies the other half: who asked about your category this week, what they said, how your voice sounds, and where the post should go.

Ship the feature, then ask the same agent which threads from the last seven days were about the problem it solves, and let it draft the replies.

What can Cursor do with Oetzi connected?

• Read the inbox. Every match arrives scored 0 to 100 for buying intent, with the reason written out and a tag for who is talking.
• Draft two reply variants in your voice for any conversation, held for your approval. Reply checks on your own posts run two and eight hours after publishing.
• Run a share of model check across ChatGPT, Perplexity, Gemini, Grok, and Google AI, and read back the threads and pages those answers cited.
• Draft and publish a post, with the day's trend research folded in, through your connected Postiz account.
• Pull the people who engaged with a post and record outreach so nobody hears from you twice.

How do I add Oetzi to Cursor?

Mint a workspace API key in Settings, then add Oetzi as an HTTP MCP server in Cursor with that key as a bearer token. The configuration block, the endpoint, the tool list, and the error codes are in the MCP docs. Copy it from there, because it is generated from the definitions the server ships.

If you would rather script it than chat with it, the same capabilities run from a REST API and a CLI, covered on the API and CLI page.

What are the guardrails?

Keys carry scopes: read, write, spend, and post. A read key turns Cursor into a reader of your market and nothing more. Add spend when you want sweeps and drafts. Add post only when you want it publishing.

Every paid capability also returns a quote first. The agent sees the price in credits, and the work runs only when the call comes back with approved set to true and a max_credits cap. Anything over the cap does not start. Reads, events, and webhooks are free, so only metered work bills.

Questions

Can I publish a post from Cursor?

Yes, with a key that carries post scope and a draft you approved. Posts go out through your own Postiz account to its 30 plus platforms. Publishing is free. The drafting run is one credit, and edits, rewrites, and regenerations of that draft cost nothing.

Will the agent post something without asking?

Not unless you gave it a post-scoped key and approved the draft. Approval is enforced on the server, not in the client, so no prompt wording gets around it. Autopilot for scheduled posts is a separate setting and is off by default.

What does it cost to try from Cursor?

A free account has 10 credits, no card, and they never expire. Connecting the server and reading the inbox are free. A sweep of one signal across every platform is 2 credits and refunds when nothing qualifies, and a drafting run is one credit.