Every new chat starts from zero. You re-explain your role, re-paste the style guide, re-upload the research you already summarized last week.

Open Context is where that context lives instead. You save it once, and assistants you’ve authorized retrieve it over MCP — the connector standard Claude, ChatGPT, and Cursor already speak.

Three building blocks

Everything in Open Context is one of three things.

Profile — Who you are and how you work: role, priorities, tone, domains, routines. This is what an assistant loads so it stops guessing.

Library item — Anything you saved for later: text, media, bookmarks, drafts, snippets, decisions. Private to you unless you share one publicly.

Space — A shared shelf for a team or project. Members contribute items out of their own Libraries, and everyone’s assistant can search the whole thing.

Who it’s for

Individuals who want coherent answers without re-uploading a resume, a style guide, and a research dump every session.

Teams aligning their assistants on the same terminology, specs, and decisions instead of passing copy-pasted context around.

No coding required for either.

How it works in practice

1. Capture. Save things as you go — the Chrome extension is fastest, or paste directly into the web app.
2. Tag. Tags and categories keep retrieval precise as your Library grows. Open Context suggests them for you.
3. Connect. Link your assistant once — Claude, ChatGPT, or any MCP client.
4. Retrieve. Your assistant calls tools like get_profile and search_library when your saved context is relevant, not on every unrelated question.

Open Context retrieves what you’ve put into it, so the first step matters most: save a handful of things before judging the results.

Next: Get started for the short path, then Profiles to shape what assistants read first.