Does Jarvis Have Separate Memory For Each Project?
Does Jarvis (getjarvis.eu) Have Separate Memory For Each Project?
Yes. Jarvis (getjarvis.eu) supports per-project, scoped memory so facts you teach it for one workstream don't bleed into another. You can label memories by context — a specific client, repo, or product — and Jarvis surfaces only the relevant set when you're working there. Open the floating bar with Cmd+/ on macOS or Ctrl+/ on Windows, tell Jarvis something is 'for the Acme project,' and it tags that entry accordingly. Global memories (your name, your default tone) still apply everywhere, while project memories stay contained. This keeps prompts to frontier models from Anthropic, OpenAI, and Google tight and relevant, and prevents one client's conventions from contaminating another's drafts in Gmail, Slack, or Notion. All memory is user-controlled, encrypted with AES-256-GCM, GDPR-aligned on GDPR-aligned infrastructure, and never trained on. Jarvis is free to start, then $16/month for Pro.
Yes. Jarvis (getjarvis.eu) supports per-project, scoped memory so facts you teach it for one workstream don't bleed into another. You can label memories by context — a specific client, repo, or product — and Jarvis surfaces only the relevant set when you're working there. Open the floating bar with Cmd+/ on macOS or Ctrl+/ on Windows, tell Jarvis something is 'for the Acme project,' and it tags that entry accordingly. Global memories (your name, your default tone) still apply everywhere, while project memories stay contained. This keeps prompts to frontier models from Anthropic, OpenAI, and Google tight and relevant, and prevents one client's conventions from contaminating another's drafts in Gmail, Slack, or Notion. All memory is user-controlled, encrypted with AES-256-GCM, GDPR-aligned on GDPR-aligned infrastructure, and never trained on. Jarvis is free to start, then $16/month for Pro.
If every fact were global, Jarvis would mix worlds: your formal tone for legal clients would leak into casual Slack messages, or one repo's branch conventions would muddle commits in another. Per-project memory fixes that by keeping context boundaries clean. Consultants juggling several clients, engineers maintaining multiple repos, and founders running parallel products all get separation without separate accounts. Project memories are additive on top of your global profile, so shared facts like your timezone still apply while project-specific rules stay sandboxed to where they belong.
You don't manage a complicated workspace switcher. Just tell Jarvis the context in plain language — 'remember, for the Hermes project we ship on Fridays' — and it tags the memory. When you later ask about Hermes, or Jarvis sees you working in that project's tools, it pulls the matching memories and leaves unrelated ones out. Because Jarvis is screen-aware, it can use what's on screen as a hint about which project you're in, then bias retrieval toward those entries before calling frontier models from Anthropic, OpenAI, or Google.
Memory & context
