Skip to content

How it works

Runs

A run is one round of asking every active prompt to every AI model in your schedule. It is what your wallet pays for. Buttons and rows such as Run now and Each run start or price a run.

The Run now dialog: the prompts times AI models line and the price The Run now dialog: the prompts times AI models line and the price

What happens in a run

Each run does the same work, in this order:

  1. Asks each active prompt once to each AI model in your model set, the way a customer types it.
  2. Records every answer word for word, with the links it gives.
  3. Reads up to 8 pages of your website as raw HTML, the way AI crawlers read it.
  4. Runs the website checks on those pages.
  5. Counts who each answer names and works out your numbers: Visibility, Share of voice, Position, Tone and the rest.
  6. Builds your report and your To-do list, with your fix suggestions.

The brand's first run writes its fix suggestions. Every later run keeps them and rebuilds the ready-to-paste files from your site, free.

When your site has changed since they were written, a note in your bell says so. See Improve.

A typical run takes 3 to 6 minutes. AI models read from the real product take a minute or two per answer, so they finish last. See AI models.

How it's calculated

Run price = your active prompts on every AI model in the set + the fixed part of each run

Pro takes 10% off the AI models and Hyper takes 20%. See AI models.

The fixed part is the website visit, the website checks and the answer scoring. Budget & Spending shows it as Fixed part of each run. It costs the same however many prompts you track, in every model set.

The brand's first run also writes its fix suggestions, at the Fix suggestions price, once. Later runs keep them and don't pay for them again.

If that writing fails on our side, nothing is charged and the next run tries again. After 2 failures in a row, runs stop trying.

What counts:

  • Only active prompts are asked and priced. A paused prompt sits out every run.
  • Each prompt is asked once per AI model per run.
  • If your site blocked automatic reading last time, each visit adds an unblocking charge. The price shows it as its own line.

The price card has every rate.

Example

A brand tracks 25 active prompts on the Standard set, which asks 4 AI models.

25 prompts × 4 AI models + the fixed part = $1.13 a run

The run gets 100 answers in all: 25 prompts to each of the 4 AI models. On a weekly schedule, Expected per month shows about $4.91, or $5.16 before the first run. A month holds about 4.35 weekly runs.

The brand's first run adds its fix suggestions, $0.25 once. The same 25 prompts cost $2.66 a run on Pro and $5.38 on Hyper.

Kinds of run

Run How it starts What it asks
Scheduled run Your schedule, at the frequency you picked Every active prompt, every AI model in the set
Run now The button on Prompts or Budget & Spending The same as a scheduled run. It takes the place of the next one.
First run Finish set up in setup The same as a scheduled run, plus the brand's fix suggestions
Run new prompts only The button on Prompts or Budget & Spending Only prompts that have never been asked. Your site isn't read again.
Pitch Start pitch in Prospecting, on agency accounts Up to 25 prompts, asked of ChatGPT, Perplexity and Google AI Overviews
Bulk run Start scanning on the Bulk run page 3 prompts per site on a list, asked of Perplexity
Free website check The free website check form on the remake website, with no account No AI model. It reads the site and runs the website checks.
Sign-up run Signing up from the free website check 3 prompts written from the site, asked of ChatGPT only, on us. Once per account, website and email.

See Run now, Schedules, Prospecting and pipeline and Bulk runs.

The free website check allows one check per site every 24 hours, and 3 different sites from one internet address every 24 hours. The sign-up run is free too. See Create your account.

A brand with no country yet doesn't run. Run now opens Pick your country first, and scheduled runs wait until you pick one. See Brand settings.

Run states

State What it means
Queued Waiting for a free slot. Nothing has been asked yet.
Running Asking the AI models and reading your site
Done Every step finished. Your numbers are updated.
Failed The run stopped. You pay only for the answers that came back.

Up to 3 runs can be queued or running per account at a time, and 1 per brand. A run you start past that is refused, and nothing is charged. A scheduled run that comes due then waits and starts in turn as other runs finish.

When you pay, and refunds

You pay as the run goes, one answer at a time. Nothing is paid in advance.

  • An answer that fails on our side is refunded at once.
  • If your wallet runs out mid-run, the run stops asking. You keep the answers that came back and pay for those only.
  • If you stop a run yourself, by deleting the brand mid-run, what already ran stays charged.
  • If your own site doesn't answer, the run still asks the AI models, and those answers are charged.

Each charge is a Run usage row in your wallet. A refund is a Refunded row.

Where to find it

  • Budget & Spending: Each run and Expected per month for your schedule. Open Budget & Spending
  • Run now: the button shows the price before you start.
  • Wallet: History lists each run's charges and refunds, by day.
  • Overview: the header shows the last run and the next one.

FAQ

Does a failed run cost me anything?

Only what ran. Answers that fail on our side are refunded. Answers that came back before the failure stay charged.

Why did my run cost less than the quote?

Answers that fail on our side are refunded as the run goes. Fix suggestions that fail to write on our side come back too, and after 2 failed tries runs stop writing them on their own. So the charge can come in under the quote, never over.

Do paused prompts cost anything?

No. A paused prompt is not asked and not priced. It keeps its history.

Can I run several brands at once?

Yes, up to 3 runs at a time per account. Start the next one when one finishes. A scheduled run that comes due at the limit waits and starts in turn as other runs finish.