Does anybody need multi-llm-multi-user shared context mcp?
Idea is this: create a project once and then decisions, open questions, instructions, files and every teammate’s AI (Claude, ChatGPT, Cursor, Gemini, whatever) works from the same context. No more re-explaining the project five times because everyone’s using a different AI. Built team-first from day one, set up entirely from inside your AI client, works with anything that speaks MCP.
You could connect it to chatgpt and say add xyz@gmail.com and they’d get an email with authentication link and connect their choice of LLM with the same project.
Would be cool for hackathon collaboration, school work collabs etc.
What do you think?