What the morning run actually does - and why your Claude limit is gone before you start work

You sit down to work, ask Otto one thing, and Claude tells you the usage window is spent. A performer wrote in with exactly that, opening with "I'm not sure what is in the /morning command" - all he knew was that the sweep took a large chunk and there was nothing left for the work he actually had to get done. If your version is "the morning run uses all my tokens", "my Claude limit is gone before I start work" or "it eats my usage window every morning", you are in the right place. The morning sweep is real, it is the heaviest run of your day, and you can move it, pause it, or slim it down.

What is in the morning command

The morning command is the daily sweep: one big pass over your leads and your inbox that

• checks what came in overnight,
• works out which follow-ups are due,
• drafts whatever needs drafting,
• and lays out your day and your approval queue.

So when people ask what does /morning do, that is the answer. It is a full working pass that has to read your pipeline before it can decide any of those four things, not a report being fetched from somewhere. In the terminal it is a plain request: /booked:run Run my morning sweep. In the app it usually lives as a scheduled task at a set time, which is how it empties your window before you have typed a word.

The Today tab is not the sweep

The Today tab is the panel where you read the result. The morning sweep is the run that produces it. Reading Today costs you nothing on your Claude plan, because it reads files already on your computer. The sweep is the part that spends. Our older guide Coming from the terminal version: your old commands in the new app says the morning command became the Today briefing and no command is needed. That is true about where you read the result. It is not true about the cost.

Why it is the heaviest run of the day

Working out which follow-ups are due and drafting them in your voice means pulling real records, not a summary. On top of that, every scheduled run carries pipeline machinery that a one-tap Play never does: triaging the whole pipeline, truing up gig records, keeping the workspace lean. That is why a sweep costs many times what a Play costs.

The app says so itself. Start Autopilot with the full sweep left in the task box and the note underneath reads: "Full pipeline sweep: Otto triages your whole pipeline, trues up records, and builds lead lists. This is the most expensive kind of run on your Claude plan. To spend less, describe just the one thing you need instead."

None of this is a Booked Solid meter, and there is no charge per token from us. Booked Solid runs on your own Claude account, so on Claude Pro your monthly fee is the entire spend, and what runs out is that plan's usage window, which refills on its own. Your Claude usage limit: why it runs out, what uses it, and how to make it last covers spending in general, but the headline belongs here: the morning sweep is the biggest single spender of your day.

Can I move the morning run to another time

Yes, and it is the first thing to try. Read the reason before you pick the new time. Moving the sweep does not make it cheaper - it costs the same at 4am as at 9am. The point is to spend the window at an hour you are not working, so it has time to recover before you sit down. Move it to 8:55 for a 9:00 start and you gain nothing.

1. Open Settings and stay on the General tab.
2. Scroll to Runs on its own, the section the app describes as "Things I do for you at a set time while the app is open". The same standing work sits under Automations in the left sidebar, where the Scheduled tasks row has a Manage button that jumps straight to it.
3. Each row reads Daily at your chosen time, plus an honest line about the last run: ran today at 08:03, last ran with a date, or last try didn't finish.
4. You cannot edit the time on a row that already exists. Choose the same preset in the dropdown, set the new time in the time picker beside it, press Add, then press Remove on the old row. In that order, so you are never left with no sweep at all.

Scheduled tasks run daily, and the list has no weekday-only option. For Monday to Friday, use the pause toggle below.

How do I pause the morning run and run it on demand instead

If you do not get new leads every morning, a daily sweep pays full price for a mostly empty inbox. Pausing is the better pattern for most people.

1. In Settings > General > Runs on its own, find the sweep's row.
2. Flip its on-off toggle off. The schedule stays in the list with everything you set. It stops firing, and nothing is deleted.
3. On a morning you do want it, press Run now on the same row. The app answers "Running - watch the Activity view."

Remove deletes the schedule for good. Use the toggle unless you are certain you are finished with it.

The catch that undoes both fixes: the app has to be open

Autopilot and every scheduled task run only while Booked Solid OS is open. Quit the app and the sweep does not fire. There is no background service, and nothing runs overnight on a closed laptop.

A missed run is not skipped, though, and this is what catches people out. The scheduler asks whether the clock has reached the task's time today, not whether it is exactly that time. So a sweep set for 4:00 on a laptop you close at night fires when you open the lid at 8:30 - precisely the moment you were trying to protect. A window that drains at 8:30 anyway looks like proof that moving the sweep was useless advice. It was not. The machine was asleep. A time you are not working has to be a time the computer is awake and the app is open: a lunch break, dinner, or a machine you leave running. For a genuinely unattended overnight sweep, read Keeping Booked Solid awake: a spare computer, or your own VPS.

If the sweep is still eating a huge chunk

Recent versions load only the slice each run needs instead of reading your whole brain every time. So a sweep that is still enormous is a signal: something specific in your setup is being read whole every single day.

1. Check what the last run reported on its row under Runs on its own, then open that run in the Activity view to see the steps it took.
2. Click Advanced on the right of the top bar and open Token history under History. Set the range to 1 day or 7 days and leave it on Whole runs to see which run cost what, then switch to Moment by moment to see where inside the sweep the tokens went. A run that hit the wall is marked hit plan limit.
3. If one sweep is wildly out of line with the rest, press Flag on its row. That saves a support bundle carrying the run's full token ledger: counts, times and task names only, unless you tick the box to include the conversation.

The usual culprit is a BookingHQ that is not lean. Closed deals, lost leads and past one-off gigs that were never archived are still live records, so every morning pays to re-read history that will never book again. Ask Otto to archive them. Archiving is reversible and never deletes: the record moves into archive/ and leaves a one-line roll-up behind. For an annual or recurring client, ask him to park it with a wake date instead, so it sleeps until its window opens.

Three dead ends

• Raising Thoroughness. It is a natural first thought, and it is backwards. Advanced > Thoroughness sits on Balanced on a fresh install, which is right for a sweep. Most thorough is for pricing, outreach, and locking in a booking. Put the sweep on it and you empty the window faster, not slower.
• Hunting for a Booked Solid usage or billing setting. There is not one to find. Nothing we change buys back a Claude usage window, because we never metered it.
• Re-running a sweep that died. When a scheduled run has already done expensive work and then fails to confirm it finished, the app deliberately stops trying it again that day, and its row reads last try didn't finish - done trying today. That is protection, not a fault. Read what it already did before firing it again, or you pay for the whole pass twice.

When the plan is honestly too small

Sometimes the sweep is fine and the pattern is the problem. Two things help more than anything else in the app: keep pure talk-it-out thinking in plain Claude, where your business brain and connector tools are not loaded and it burns far slower, and start a fresh conversation per topic rather than one endless thread that re-reads its whole history on every message. If you run a daily sweep and chat all day, Claude Pro may be the wrong size for you, and that is a decision inside your Claude account at claude.ai.

Still eating the window?

If the sweep still takes a huge chunk after you have archived what is done, moved or paused the schedule, and looked at Token history, we would rather see the numbers than guess. Press Flag on that run in Token history, then email bookedsolid@kivimedia.freshdesk.com with the bundle, which Claude plan you are on, and roughly how many active leads you carry. A run's token ledger shows exactly what it loaded, and that is the difference between us guessing and us fixing it.