Run frequency

Every prompt has a schedule that decides how often AmICited asks it again. AI answers change constantly, so re-running a prompt is how you see a trend instead of a snapshot. Each run also costs credits, so the schedule is the main lever on your monthly spend.

Available schedules#

The interval is counted from the prompt’s last successful run. A new prompt is picked up within minutes of being added. If a run fails, it is retried automatically after about five minutes. one_time prompts run once and then keep their results without re-running, which suits one-off audits.

Setting the schedule#

• Setup wizard: the Frequency field on step 2 applies to every prompt created in the wizard.
• Add prompts dialog: the Schedule field applies to the whole batch. The dialog notes that more frequent cadences use more credits.
• One prompt: Edit prompt on the Prompts page, or the schedule control on the prompt’s detail page.
• Many prompts: select rows on the Prompts page and use Frequency: in the bulk bar.
• Right now: Process now in the bulk bar runs the selected prompts immediately, without waiting for their next scheduled run. It costs the same as a scheduled run.

The Prompts page can be filtered by cadence, which makes it easy to review which prompts run most often.

The credit math#

Each engine a prompt runs on returns one AI response, and a response on a standard engine costs 0.015 credits. So:

• Engines is the number of engines the prompt runs on: 5 on Starter and Pro by default, up to 8 on Premium and Enterprise. A prompt restricted to fewer engines costs less. See AI engines.
• Premium engines (Claude, Grok, DeepSeek) are metered at the vendor’s cost and never less than 0.015 credits, so treat 0.015 as the floor for them.
• Keyword data lookups (0.005 credits per keyword, free when cached) and other features are metered separately. See credits.

Cost of one prompt on five engines#

One hourly prompt costs as much as 24 daily prompts. Reserve sub-daily schedules for a handful of prompts where you need to see changes within hours, such as during a launch or right after publishing a page.

Worked examples#

Starter, 20 prompts daily. 20 × 5 engines × 30 runs = 3,000 responses. 3,000 × 0.015 = 45 credits a month, which fits Starter’s 50 monthly credits with 5 to spare.

Starter, mixed cadence. 10 core prompts daily (10 × 2.25 = 22.5) plus 80 long-tail prompts weekly (80 × 0.32 = 25.7) comes to about 48 credits a month. You track 90 prompts instead of 20 for the same budget.

Pro, 40 prompts daily plus 100 weekly. 40 × 2.25 = 90, plus 100 × 0.32 = 32, is about 122 credits: just over Pro’s 120. Move 10 daily prompts to weekly and you are back under.

Premium, one prompt on all 8 engines daily. 8 responses a day × 0.015 = 0.12 credits a day, or at least 3.6 credits per 30 days. The three premium-engine responses may cost more than the floor, depending on how much each vendor’s search and tokens cost for that prompt.

Choosing a cadence#

• daily is the right default for your core category and comparison prompts. It gives a clean day-by-day trend and enough runs to smooth out the natural variation between answers.
• weekly suits long-tail, branded and informational prompts where you care about the direction over a quarter, not the day-to-day swing.
• hourly and every_3_hours are for short, deliberate experiments. Switch them back when the experiment ends.
• monthly and yearly keep a large reference set alive at almost no cost.

More runs do not make a single answer more accurate, but they do make the averages steadier. A prompt with 30 runs a month tells you far more about its real citation rate than one with 4.

Related#