About
One person, one product, no write access.
LinkIntel is not a platform, a suite or a company with departments. It is a small service that gives the assistant you already write with evidence for your next post: your own X results, X's ranking code, and where each came from.
Who runs it
Marcos Placona
I am a developer relations consultant based in London. I write posts most weeks, and I draft them with an AI assistant, which is how I ended up needing this: the assistant was writing from a general idea of what performs, not from posts I knew had worked.
LinkIntel is operated by DevRel Bridge Ltd, a company registered in England and Wales (company number 15979110), registered office 128 City Road, London, EC1V 2NX. There is no staff behind that name. If you email support, I am the one who reads it, and I am also the one who can change the thing you are writing about.
Why it exists
The dashboard was the wrong shape. So it went.
LinkIntel started as a dashboard for LinkedIn Page analytics. It worked, and I used it, and it changed nothing about what I posted. The numbers were in one place, my writing happened in another, and the gap between those two places was where all the good intentions went to die.
Then I noticed I had stopped drafting posts in a text editor: I was drafting them by talking to an assistant. The dashboard was competing with a habit it could never win against. So it was retired in August 2026 and LinkIntel was rebuilt as an MCP server, living inside the conversation where the writing already happens.
It starts with X because X is where the evidence can be read for any public post you name, including other people’s. Hand your assistant the posts you want to learn from and it gets their real public performance, with a link back to each one, instead of guessing. LinkedIn is in closed beta and not open to new accounts. If you already connected LinkedIn analytics, it stays available on your account page.
Plainly
What it is, and what it is not.
An assistant acting on your behalf should know exactly where the edges are, so the boundary matters more than the pitch does.
It is
- Your recent original X posts and their metrics once you connect X, plus public X posts you choose as references, each with source links
- X's open-source ranking weights, read daily
- A cited content brief your assistant reads before it drafts, with the limits of the evidence stated
- Delivered as an MCP server, so it works inside Claude and other assistants you already write with
- One plan, $39 USD a month, with the amount shown before anything is charged
It is not
- Not a scraper. It reads your own posts only if you connect X, looks up only public posts you hand it or a list you agreed to, and never searches X
- Not able to post, comment, react, message or follow as you. It holds no write permission at all
- Not a virality predictor. Someone else's reach is evidence to test against, not a forecast of yours
- Not a place your data gets sold, shared or used to train a model
What it needs from X
Nothing to get started. Never your password.
Public reference posts need no sign-in. If you choose to connect X, you approve a read-only consent screen on X and never give LinkIntel your password. Disconnecting deletes the stored snapshots. Existing LinkedIn analytics customers grant a single read-only permission, r_member_postAnalytics, which they can take back on LinkedIn’s permitted services page.
If that sounds like the shape of thing you wanted, it is right here.
Sign in with Google, confirm $39 a month, and connect the assistant you already use. Everything after that happens in the conversation.