Free tool · All tools
Know what a run costs before it repeats.
A one-off prompt can cost whatever it costs. Recurring work multiplies by its schedule, so estimate the token math first — per run, per month, against a published allowance.
The rate presets are illustrative tiers, not live provider prices. Enter the current rates for your model; token math stays the same.
Estimated AI spend
- Per run
- $0.23
- Per month
- $6.75
Published plan allowances
| Plan | AI / month | Run starts | Covers this |
|---|---|---|---|
| Free$0 | $0.50 | 10 | No |
| Starter$10/mo | $2.00 | 50 | No |
| Pro$39/mo | $10.00 | 250 | Yes |
| Scale$99/mo | $25.00 | 1000 | Yes |
This estimate fits the Pro plan's published AI allowance and run starts.
At a limit, only the affected new work pauses. There are no automatic overages or surprise bills.
Compare monthly plansQuestions
How is the cost of an AI run calculated?
Cost per run is input tokens divided by one million times the input rate, plus output tokens divided by one million times the output rate. Multiply by runs per month for the monthly figure. That is the whole model — no hidden multipliers.
Why are input and output tokens priced separately?
Providers meter them separately, and generated output typically costs several times more per token than input. A run that reads a lot and writes a little prices very differently from one that writes long reports.
Are the rate presets real model prices?
Model rates change frequently, so the presets are illustrative tiers rather than claims about any provider's current pricing. Enter the rates from your provider's price list; the arithmetic is the same.
What happens when an estimate exceeds a plan allowance?
Each plan publishes a monthly AI allowance and a number of run starts. When new work would exceed an allowance, only that new work pauses — data and configuration stay intact, and there are no automatic overages.
Start with Free