Cursor ↔ Claude on shared context.

The pain

• —You started a feature in Cursor with 30 minutes of context-priming. You hit a thinking dead-end, want a second opinion from Claude.
• —You paste the conversation into Claude. Half the context is gone — Claude doesn't have your codebase, your design decisions, your customer feedback.
• —You re-prime Claude for 20 minutes. Get unstuck. Want to go back to Cursor. Repeat the priming there.
• —Most of your AI time is spent re-explaining what you already explained an hour ago.

What you do in Memory.Wiki

1. 1
   Build the project hub once

   Capture the spec, recent decisions, customer feedback, design constraints into Memory.Wiki docs. Group them into a Bundle named after the project. That bundle has a permanent URL.
2. 2
   Paste the bundle URL when priming

   Cursor: drop the URL into .cursorrules or the project README. Claude: paste it in the system prompt or first message. Both AIs fetch the same markdown.
3. 3
   Switch freely

   Hit a wall in Cursor, ask Claude — they're both reading the same hub. The context is the URL, not whichever chat thread you happened to be in.
4. 4
   Keep the hub current

   When a new decision lands, save it to Memory.Wiki (Hub Chat's Save-as-doc, or the editor). Next time you paste the URL, both AIs see the updated state. No re-priming.

What you get back

• ✓Setup cost moves from "every conversation" to "once per project."
• ✓Switching between AIs is a paste, not a translation.
• ✓When a teammate joins, they paste the same URL. They onboard at AI speed.
• ✓Your context survives tool churn — Cursor → Claude → ChatGPT → next year's tool — without you having to migrate anything.

Worked example

Try it with what's on your desk right now.

No signup. Drop in your first doc and the URL is yours.

Other shapes the URL takes