Lotti Manual
Keep the facts. Make the day yours.
Not another machine for pretending everything went to plan
Lotti distinguishes the things you meant to do from the things that actually happened. A task can be important without owning your whole day. A plan can be rearranged without falsifying history. A time record can say, honestly, that the meeting ran long and the good idea arrived while you were walking home.
That separation is the point, not paperwork. It gives you a record you can trust when the week gets noisy.
The penguins, orbital habitat, sardines, and Project Waddle appear only in this manual's fictional Intergalactic Penguin Logistics demo workspace. They make the screenshots feel like one coherent story, while the controls, flows, and product behaviour shown are the real Lotti app. Your own workspace can be as ordinary—or as delightfully strange—as you like. Task cover art is optional; a good visual metaphor can make a crowded week much easier to read.
The worldscape is not starter content and is never added to your profile. It is a deterministic screenshot fixture: one silly logistics operation, recurring characters, and distinct task covers make it possible to follow a feature from the agenda to the task, project, journal, and dashboard without mistaking sample data for a product promise. The cover art is equally optional in real work: create it in Lotti when a useful, serious, or gloriously peculiar image makes the work easier to recognize.
Start where the friction is
Your life is not a backlog See how intent, plans, actual time, evidence, and reflection stay distinct—and useful.
Build a day that can breathe Turn a spoken or typed check-in into an editable, reality-aware day plan.
Give work a home Create tasks, attach context, set priorities, and turn ambiguity into next moves.
Choose a brain responsibly Pick local or cloud inference deliberately—privacy, retention, region, and energy are all part of the choice.
AI is optional. Its consequences are not.
Lotti has no inference backend and no copy of your model input. Local storage and local inference are related choices, but they are not the same promise: your logbook stays on your devices either way. With a local provider, selected material is processed by the machine or local endpoint you configure. With a cloud provider, the chosen context goes directly to that provider under your account and API key. That makes the decision yours—but it also makes the due diligence yours.
For sensitive material, run a capable model locally or choose a provider only after checking its retention, training, jurisdiction, and account settings. “Cloud” never automatically means “nothing is stored.”
AI runs in physical places. Prefer a local setup or provider whose electricity, location, and carbon claims you can actually verify. A renewable-powered route can avoid outsourcing the health and climate cost of fossil-powered compute—especially gas-turbine generation—to communities with less power to refuse it.
Agents can analyse context and draft changes, but changes to your record wait for you. The only narrow exceptions are an initial title or language for an otherwise empty task. Lotti does not turn “helpful” into “unaccountable.”
The practical rule: if you cannot find a data and energy path you are happy to stand behind, leave AI off. Lotti remains useful without it.
Create a couple of meaningful categories, one real task, and one Daily OS check-in. Add more structure only when it earns its place; you are building a record of a life, not feeding a productivity machine.