Credits measure the work the AI agents do for you: building, running and repairing tests. Everything you do yourself — editing, reviewing, reporting, admin — is free.
What uses credits, and what does not
Uses credits — agent work
|
Always free
|
Where credits actually go
The shape matters more than the numbers, because it explains the bill you see in month one versus month six.
| Create a test | One-time, per test | By far the largest share |
| Diagnose / fix | As needed | Moderate — a fraction of a rebuild |
| Run a test | Every execution | Small |
Usage is front-loaded. Most credits go into building your suite. Once it exists, running it costs far less — one new test costs roughly the same as about fifty runs. Expect heavier usage early, then a steady run-rate.
This is the single most useful thing to explain to whoever approves the spend, because the first invoice is not representative of the steady state.
Plans at a glance
|
Free $0, no time limit |
Growth $20 / month |
Scale $100 / month |
Enterprise Custom, annual |
|
| Credits included | 800K / mo | 3.2M / mo | 20M / mo | Custom pool |
| Roughly covers | 2 new tests + 100 runs | 8 new tests + 400 runs | 50 new tests + 2,500 runs | Sized to your footprint |
| Credit sharing | Team workspace | Shared team pool | Shared team pool | Pooled across all teams |
| Need more? | Upgrade anytime | Buy more anytime | Buy more anytime | Add-ons at contract rate |
| Unused credits | Reset monthly | Reset monthly | Reset monthly | Roll over for the term |
The "roughly covers" figures are illustrative. They assume a mid-length test of around 20–40 steps and an even split between building and running. Actual consumption depends on test length, complexity, and how often you regenerate or diagnose.
Current pricing is at functionize.com/pricing, and limits at functionize.com/usage-limits.
Enterprise: a pool sized to you
- Baseline your footprint — annual test executions, plus the new tests you plan to build.
- Convert it to a monthly pool, with headroom for growth.
- Share it across every team — one pool, unlimited users, no per-seat math.
- Review it together — we check actual usage with you mid-term and right-size before renewal.
Users are unlimited and included. Unused credits roll over month to month for the term, so busy months draw on quiet ones. Add-on credits are available at your contracted rate.
Running low: you are never locked out
This is the part worth knowing before it happens.
| Always keeps working | Your tests, results and data, plus manual editing, dashboards, integrations, exports and admin. |
| Pauses until credits are added | New agent work — creating, running, diagnosing — resumes once you add credits, upgrade, or reach the next cycle. |
Admins can see usage live in Studio and in Reports, so this should never be a surprise. Check it during your first month while your build rate is highest.
Getting more from every credit
These are the habits that materially change consumption.
Fix, do not rebuild. When a test breaks, use diagnose and fix rather than regenerating it. A repair is a fraction of the cost of a full rebuild, and this is the biggest single saving available. See Diagnosing a Failed Test.
Be specific up front. Clear intent and explicit verification points mean fewer regenerations. A vague prompt that needs three attempts costs three times as much as a good one. This is the practical reason the exercises spend so long on how to phrase a request.
Make small edits by hand. Inline edits and manually added steps are free. If you know exactly what needs changing, changing it yourself costs nothing.
Reuse flows as Components. Build login, navigation or checkout once and reuse it everywhere, rather than regenerating the same journey inside twenty tests. The Add component control sits in the steps pane toolbar.
Schedule with orchestrations. Run suites together, off-peak, with one consolidated report, instead of many ad-hoc individual runs.
Prune stale tests. Archive tests for retired features so scheduled runs stop consuming credits on things nobody ships any more. Worth a quarterly pass.
Common questions
Why credits instead of seats or VMs? You pay for the work agents do, not for headcount or idle infrastructure. You can add people freely without adding cost.
Will costs balloon as we scale? No. Functionize runs its own purpose-built models, so costs are not tied to third-party token pricing, and the platform gets less expensive to run as you scale.
Why did our first days use so many credits? Building tests is the heavy part, and a trial is mostly building. Once your suite exists, day-to-day consumption is mostly light, recurring runs.
Do unused credits carry over? On self-serve monthly plans they reset each cycle. On Enterprise they roll over month to month for the full term.
Sizing a plan
Two numbers decide it: how many tests you intend to build, and how often you intend to run them. Because building dominates, a team creating a hundred tests in a quarter has a very different first-quarter profile from a team running a hundred existing tests daily — even though both might describe themselves as "a hundred tests".
Work out the build number from your critical journey list, and the run number from your intended schedule. Your account team will map that to the right pool; talk to them rather than guessing.
Connecting credit consumption to business outcomes month over month is covered in What to Measure: QE Metrics and KPIs.