Firedrill pricing and credits
Source: https://firedrill.run/pricing
Start free, then choose a plan with monthly credits for your organization. One credit balance covers world preparation, simulations and standalone tools.
Prices in USD, billed monthly. Creating an account does not require a payment card. Billing shows the plan, credits and rates that apply to your organization before you confirm a subscription or purchase.
Plans
| Plan | Price | Credits each month | Builders |
|---|---|---|---|
| Free | Free | 500 | 1 |
| Starter | $29 / month | 3,000 | 1 |
| Team | $99 / month | 12,000 | 5 |
| Scale | $299 / month | 40,000 | 15 |
| Enterprise | Custom | Contracted | Contracted |
A synthetic version of your business, scoped with our team. Talk to us.
Extras
Credit top-up: $10 per 1,000 credits
Add prepaid credits on any plan. Automatic top-ups stay off until a billing owner sets a threshold, amount and monthly cap.
Additional builder: $15 per builder per month
Available on paid plans. Each additional builder adds 1,000 credits a month to the organization balance.
Credit rates
| Usage | Credits |
|---|---|
| Synthetic-user input | 3 per 1,000 tokens |
| Synthetic-user output | 15 per 1,000 tokens |
| Scoring | 0.1 per 1,000 tokens |
| Synthetic tool | 10 per tool-hour |
| World preparation | 0.05 per CPU-second |
| Custom isolated tool | 60 + 20 + 10 per vCPU-hour, GiB-hour and instance-hour |
Usage is measured as it happens and charged on the cumulative total, not rounded up per call or per hour. For example, 8,000 synthetic-user input tokens, 800 output tokens, 20,000 scoring tokens and three synthetic tools running for two minutes use 39 credits, plus any world preparation. A credit is not a fixed number of simulations.
What uses credits
Prepare a world
Preparing the selected synthetic tools and data, generating scenarios and checking that the world can run. Reusing an unchanged prepared world does not repeat the preparation charge.
Run simulations
Synthetic users and scoring during each drill. More scenarios and repetitions use more credits. Your agent runs on your machine.
Keep tools running
Synthetic tools used in drills and standalone tool sessions, for the time they stay allocated. Stopping a tool ends its runtime charge.
Your organization's balance
Monthly credits
Each plan adds its credits to one organization balance every month. Billing shows the source, any expiry and the next allocation.
Top-ups
Prepaid credits you add when you need more. No postpaid overage: new work does not create an unpaid bill.
Credit codes
Redeem an eligible credit code in Billing. It adds credits without starting a subscription or automatic top-up.
Builder seats
The organization owner assigns seats to people who build. Viewers, invitations and agent identities do not become paid seats.
Estimate, maximum budget and actual use
Before any work starts, you see the estimated credit cost and the maximum budget. The estimate is the expected use; the maximum is a ceiling, not a charge. Firedrill reserves credits for approved work, charges actual use and releases the rest.
Reading reports, opening saved evidence and individual tool calls are not charged separately. Infrastructure failures and recovery of the same request are not billable. A real simulation can use credits even when your agent's result fails a check.
Capacity and separate costs
A credit balance does not guarantee unlimited capacity or concurrency. Review the selected resources and limits before running or extending work.
Your agent's own model provider charges you separately and is not billed again by Firedrill. Running your agent on your own machine does not use hosted compute credits.
How credits work (app sign-in may be required).