About / Responsible AI

Schedules, briefs and chat replies are starting points, not decisions. Each field can be edited inline, regenerated, or discarded. Treat the assistant as a first-pass writer whose work you always review.

Generation is based on pattern matching over the text you supply. It can misread intent, over-simplify nuance, invent structure that isn't in your source, or mis-order priorities. Verify facts, dates, names and figures against the original source before sharing anything.

Avoid client identifiers, personal data, credentials, contract terms, medical or financial records, and anything under NDA. Data stays in your browser's localStorage, but a shared or unlocked device still exposes it — use Settings to clear all data when you're done.

This is a fully self-contained front-end app. There is no account, no server, and no external AI provider. 'AI' results are produced by local JavaScript — keyword scoring for the planner, extractive summarisation for research, and template-based drafting for chat. Nothing you type leaves your device.