Drop in your LinkedIn connections and they arrange themselves by what people do. Thousands of names become a few dozen clusters you can orbit, search and question in plain English.
No file yet? Here’s how to get it — it takes about ten minutes.
LinkedIn will give you a copy of your own data. No password to hand over, no browser extension, nothing that breaks when LinkedIn changes its site.
Open LinkedIn’s Get a copy of your data page (Settings → Data privacy). Pick Want something in particular? and tick Connections.
LinkedIn emails you when it is ready, usually within ten minutes. Download the archive and unzip it.
Drop Connections.csv on this page. It is read in this tab and kept in this browser, and nowhere else.
Any CSV with a name column and a headline or job title works too. You can say which column is which before anything is built.
Everyone is placed by the field they work in, and people who name the same employer are joined through it. Orbit, zoom, or isolate one field.
“Anyone in satellite imagery licensing?” is answered on your machine in milliseconds, with the query it ran shown above the answer so you can check its reasoning.
Type a name and the camera flies to them. Click anyone for their panel: role, employer, and who else you know there.
With your own Anthropic API key, Claude places employers so “any VC in SF?” works, and researches one person on the public web when you press Enrich profile.
There is no backend, no account and no analytics. This is everything that can leave this tab, in full.
Your file, the graph built from it, and every search and question that is answered locally, which without a key is all of them.
Employer names, to look up headquarters and organisation type, and the text of your questions, to read them more closely. Sent from your browser straight to Anthropic, with your key, which is kept only if you tick Remember.
One person’s name, headline and employer, so Claude can search the public web for them. Never automatically.
Never sent anywhere: email addresses or profile links. Without a key, the one request this page makes to anyone else is for the 3D library, from jsDelivr.
No build step, no dependencies, no server. Read every line, run it on your own machine, or put your own copy on GitHub Pages in a few minutes.
git clone https://github.com/asturksever/network-constellation
cd network-constellation
npm run devEverything so far works without one. With a key, Claude can place employers, read your questions more closely and research a person on the public web — each only when you ask.
Sent to Anthropic: employer names and the text of your questions. A person’s name, headline and employer only when you press Enrich profile on their panel. Never emails or profile links. Everything is saved in this browser, so nothing is paid for twice. Use a key with a spend limit.
Your export has no location and no company type, so “any VC in SF?” can’t be answered from it. This looks up every employer at once — company names only, no people — so location questions work. It is separate from enriching a person.
Which employersLines join a person to their field and to an employer named by two or more people. They are not links between the people themselves — LinkedIn doesn't share who knows whom.