Skip to main content

Your next three

Open Today and the first thing you see is three tasks, in the order Latte thinks you should take them, with one button to start the first. Not your whole list — three.

The point is to end the decision. Looking at forty open tasks to choose one costs more than the task does, and it is the moment most people close the app.

Screenshots pending

This page is written from the code, not from a running app — no environment was available to shoot it. The images below show a described placeholder instead.

  1. Open Today from the navigation rail. Up next sits at the top of the page: one task, its course, when it is due, and one line saying why it came first.

  2. Directly beneath it, Your next three lists the two that come after — a number, the title, the deadline and the same one-line reason each.

    Screenshot coming soon
    The Today page showing the Up next hero card and, directly beneath it, the card listing the two tasks that come after — each a numbered row with title, course chip, deadline, a one-line reason, and the Just start / Not now buttons

  3. Just start — 10 min begins a focus session on that task with a ten-minute goal. Ten minutes is the whole commitment; you can stop there.

  4. Not now moves that task down and brings something else up.

  5. With nothing open, the card reads Nothing queued — "Capture one thing, or connect your schedule, and it will show up here."

On the Latte desktop app

The desktop app shows the same two cards in the same order.

If the card says it couldn't load

We could not load your tasks means exactly that — "This is a loading problem, not an empty list." It is never Latte telling you that you are done.

What the order is based on​

Latte ranks the tasks you already have. It reads:

  • How soon each one is due, and how much is left between now and then.
  • How long you have before your next commitment, and how long the task looks like it takes — a ninety-minute task is a poor suggestion twenty minutes before a lecture.
  • What kind of task it is — an exam revision block and a reading are not the same weight.
  • What you just finished, so a run of similar work can continue.
  • What you keep pushing away. Two or three Not now taps on the same task and it stops being offered first — and the reason line changes to one about simply breaking the ice.

What it never looks at​

Grades. Latte has no marks to read: there is no grade, GPA or score anywhere in what the ranking sees, and the app does not ask for any. It will never tell you that a task is worth more because you are doing badly in that course, and it will never predict a result.

It also does not read your messages, and it does not rank work for anyone but you.

Why it stops at three​

Three is a hard limit, not a page size. There is no "show the rest" on these cards — your full list is one tap away under Tasks, and this is deliberately not that. See Edit and complete tasks when you want the whole picture.

What "Not now" teaches it​

Every Not now is a small signal, and it is kept: the task drops in the order, and the ranking's sense of what suits you shifts a little. Start a task and the opposite happens. This runs on your own device and carries over between sessions — you are not training anything shared, and nobody else sees it.

You can undo the effect simply by starting the task next time it comes up.

The Premium AI-assisted ranking​

On the Premium plan you can switch on an AI-assisted ranking, which re-orders the same tasks and writes its own one-line reason for each. It is a switch you turn on yourself, not a default, and the free ranking above keeps running underneath it — if the AI is unavailable, slow, or over its limit, the order simply stays as it was.

Wherever a reason line was written by the AI, it carries the AI-generated label. A reason without that label was produced by the plain rules described above.

Turning it on does not change what the ranking can see. It still has no grades, and it still never predicts one. See Text an AI wrote and Premium features.