Teach your AI to read your context, and to save what it learns.
Connected your AI to UseMyContext.ai? Two pastes finish the job. Most AI clients will not reach for your context unprompted, and none of them offer to save what a conversation taught - so you name the tool by hand at the start, and lose the good part at the end. Drop both snippets below into your AI's standing instructions once. The first makes it read your context; the second makes it offer to save what the session taught.
1. So it reads your context
Stops the "I don't know anything about you" answer. Your AI checks your profile and your documents before it answers.
Before answering anything about me, my work, or my documents, check UseMyContext first (profile + ask_docs). Never say you don't know about me without searching it.
2. So it offers to save what the chat taught
Reading is only half the loop. This one has your AI offer a save as a conversation wraps up, so what it just learned about you does not die with the chat window.
When a conversation surfaces something new and durable about me, offer to save it to my UseMyContext profile before we wrap up. I review every suggestion before it lands.
Where to paste them
Both go in the same box, one after the other.
- Claude - Settings -> Preferences (or a Project's instructions).
- ChatGPT - Settings -> Personalization -> Custom Instructions.
- Cursor - a rules file (or our Cursor plugin, which applies it automatically).
Both snippets work in any client you've connected over MCP, and neither changes anything about your control: your AI still only reads what you've approved, a save arrives as a suggestion you accept or ignore on your Profile page, and every access leaves a record.
Not connected yet? Connecting takes one click from Claude's directory or ChatGPT's plugin directory - the connect guide covers every client.