How Do I Scope A Memory To One Client Or Workspace?

Tell Jarvis (getjarvis.eu) the context when you save the fact. Say something like 'for the Acme account, the main contact is Dana and we bill monthly,' and Jarvis tags that memory to Acme so it only surfaces when you're working on Acme — not when you're drafting for another client. Global facts (your name, your timezone) still apply everywhere; scoped facts stay contained. Because Jarvis is screen-aware, it can also infer which client you're in from what's on screen and bias recall accordingly. Open it with Cmd+/ on macOS or Ctrl+/ on Windows. Scoped memories work across connectors like Gmail, Slack, and Notion, are routed to frontier models from Anthropic, OpenAI, and Google as needed, and are encrypted, GDPR-aligned, and never trained on. Jarvis is free to start, then $16/month for Pro.

You don't configure folders or rules. Just include the context in your instruction — a client name, a project, a workspace — and Jarvis attaches the memory to that scope. 'Remember for the Hermes project we use Linear, not Jira' becomes a Hermes-scoped fact. The next time Hermes comes up, Jarvis pulls it; when you switch to a different project, that fact stays out of the way. This keeps each engagement's conventions clean and prevents cross-contamination in your drafts and actions.

At query time, Jarvis combines a few signals: the scope you mention, the topic of your request, and what it sees on screen. If you're clearly working in Acme's Notion space or emailing an Acme contact in Gmail, it weights Acme-scoped memories higher before sending context to frontier models from Anthropic, OpenAI, or Google. You can always be explicit — 'use my Acme settings' — to force a scope. The aim is relevant recall without you micromanaging which memories apply.

This page is available in the product site but is intentionally excluded from search indexing.

Memory & context