Ask the way you remember
By person, project, decision, or roughly when — not only by exact keyword. Hybrid search runs full-text and vector retrieval together, so a half-remembered phrase and a fuzzy description both land.
Contextual recall
Ask about your week without pasting the backstory
One question across mail, chat, docs and calendar, answered from your own history with the source and time attached to every hit.
A general assistant starts every session empty, so you spend the first few minutes re-explaining your life before you can ask anything useful. ShogunAI starts from the memory it already built, which means the question can be the short one you actually had.
One question, every connected tool
Mail, chat, documents and calendar answered from a single search instead of four. What you authorized is in scope; what you did not stays out.
Answers you can check
Every hit carries its source and its time. You get "Tuesday 14:20, in the design doc" and can open it, rather than trusting a paragraph with nothing behind it.
Confidence is visible
Each piece of state the system holds — people, projects, commitments, open loops — carries a confidence score. A low-confidence read surfaces as a question, never as a fact inside something you are about to send.
From context to a useful next step
A general assistant starts every session empty, so you spend the first few minutes re-explaining your life before you can ask anything useful. ShogunAI starts from the memory it already built, which means the question can be the short one you actually had.
Ask
Describe what you need in ordinary language, even if you only remember part of it.
Retrieve
Recall searches the warm tier — the last month in full detail — and returns hits with provenance attached.
Continue
Turn the answer into a brief, a reply, or a next action without starting the context over.
What this changes
Spend less time rebuilding context
- Walk into a meeting already knowing where it left off
- Recover a decision after two weeks away
- Stop searching four apps for one thread
- Draft from facts instead of from memory
Frequently asked questions
Clear answers before you start
How is this different from search?
Search expects the exact words and one app at a time. Recall retrieves by meaning across the tools you connected, and attaches the source and time to each result so the answer is checkable.
Does it work offline?
Yes for recall. Indexing and search run on your Mac, so finding things needs no network. Generation — drafting a reply, reasoning through a hard question — is the part that needs a model.
What if it is not sure?
Low-confidence state is hedged rather than asserted. It comes back as "you may not have sent Mika the revised numbers", not as a sentence quietly hardened inside a draft.
Which plan includes it?
Standard. Capture, recall, the daily brief and the first-layer read-only connections are all in the lower plan.
Explore all Features
See how ShogunAI connects private memory, recall, and execution.
ShogunAI Command Layer
Every model you use starts from zero. Give it your week.
Connect once, keep one private memory on your Mac, and let ShogunAI finish the work from there.
Joining the early-access waitlist is completely free.
macOS · 7-day full trial · your memory stays on your device.