Skip to content

Else documentation

How Else handles your work — what it costs, what it will and won't do on its own, and how to stay in control.

Else is something you delegate to. You describe a piece of work; it makes a plan, does the work, and hands back a result — stopping to check with you before anything leaves the app, and telling you what it actually did.

These pages are about how that works in practice: how to hand a job over so it comes back finished, what bounds you can put on it, and where the record of what happened lives.

What Else is — the idea in full: what it does inside a run, what it isn’t, and the three points where you stay in control.

Your first task — one real delegation end to end, from the sentence you type to the file you download.

Writing a good request — the four things that decide whether what comes back is usable.

Giving it materials — the three ways to hand over a file, what it accepts, the size limits, and the one difference between a chat and a handed-off task.

Reviewing what comes back — how to tell finished from stalled, where the deliverables are, and what to do when the work came back wrong.

How billing works — prepaid, metered, no subscription. The five things that can cost money, and how each is measured.

Model prices — the per-million-token price of every model you can pick, and which model each tier runs by default.

Reading a receipt — every field on a receipt, what it means, and the two places receipts appear.

Topping up — how to add money, what happens the first time, and what each failure message means.

When a run fails — what you are charged when nothing useful came back, what you keep, and why there is no refund.

Spending limits — the number you set is read on work you hand off, and not on a chat. Exactly what it does, and what it does not.

When it checks with you — where it stops and waits for a decision, how to widen or narrow that, and why waiting costs you nothing.

Stopping a run — what survives, what is thrown away, and what happens to money already spent.

What it won’t do — the limits it actually has, split into the ones enforced mechanically and the ones written into its rules.

Add-ons — the 13 built-in abilities every run already has, how to narrow them for a single run, and how to teach it something of your own.

Connecting your apps — what it can do inside the accounts you already use, how your credentials are stored, and which actions stop for your approval no matter what.

Something went wrong — start here with a symptom rather than a question; each entry names the real cause and the page that fixes it.

It stopped before it finished — four different things end a run and they look identical. How to tell which one you hit.

It cost more than I expected — six checks, in the order worth doing them.

The result looks wrong — the signals that separate a checkable claim from an invented one.

A connected app isn’t being used — connecting an app doesn’t put it into a run, and the home screen can’t add one.


The product itself is at elsedoes.com, and Updates is the record of what changed in it, one report per week.