Documentation
Back
  • Plans, limits and what a scan costs
  • What Citerva measures
    • Setting up your first project
  • Running scans and reading the evidence
  • How every number is calculated
  • Project settings, pausing and archiving
  • Workspaces, members and seats
  • Competitors and topics
  • Scheduling scans
  • Recommendations, and what they are not

Search docs

Find a documentation page or section by title, description or heading.

Scheduling scans

How to run a project automatically, what a schedule costs, and why one scan is never the answer.

One scan is a sample. AI assistants are not deterministic — ask the same question twice and you can get two different answers, with different brands named. A single scan tells you what happened once.

The trend across several scans is the thing worth reading, and a schedule is how you get one without remembering to.

Setting one up

On a project's Scans page, under Schedule: pick a cadence, an hour, and the providers to ask. Save.

Times are UTC, always. A schedule set for 09:00 runs at 09:00 UTC whatever your clock says, and does not move when your country changes its clocks — which matters, because a scan that shifted by an hour twice a year would put a step in your trend line that has nothing to do with your visibility.

What each plan can schedule

PlanCadences
ProWeekly, monthly
GrowthDaily, weekly, monthly

A scheduled scan spends the same allowance

This is the part worth reading twice.

A scheduled scan costs exactly what the same scan costs when you press the button: one provider call per approved prompt, per selected provider. The difference is that nobody is watching when it happens.

So a daily schedule over 50 prompts and 4 providers is 200 calls a day, which is 6,000 a month — the entire Growth allowance. Scans are capped too: Growth allows 40 a month across the whole workspace, and one daily schedule uses about 30 of them. Before saving a schedule, multiply it out.

If a scheduled run would take the account past its monthly allowance, it is refused rather than run. No partial scan, no overspend. The schedule stays active and tries again at its next slot, so the month rolling over is enough to resume it. The Scans page shows a skipped run as Last run skipped, with the reason.

It always asks the current questions

A schedule is attached to the project, not to a frozen copy of your prompts. Each run asks the approved prompts in the newest prompt set at that moment. After you regenerate a set, its prompts wait for your review, and a scheduled run with nothing approved is skipped rather than asking unreviewed questions.

That means approving a new set changes what every future run measures — which is usually what you want, and is also why the trend chart marks a break at that point rather than drawing a smooth line across it. Two scans that asked different questions are not two points on the same series.

Pausing the project stops the schedule too

A schedule belongs to a project, so pausing the project stops its scheduled runs as surely as pausing the schedule does. The schedule is not cleared — it resumes at its next slot once the project is active again, rather than firing everything it missed.

Pausing and removing

A schedule can be paused and resumed; paused means nothing runs, and the project keeps everything it has already measured. Removing one asks for confirmation first.

Pausing a schedule does not pause the project — you can still scan by hand.