Plan tier limits

This page is the canonical reference for what each plan tier includes. Numbers come from backend/app/config/tier_catalog.py, which is the single source of truth the API itself reads at runtime.

Note

For the why behind the numbers (why BYOK is paid, how Free is fenced), see Plan philosophy. For upgrade and downgrade mechanics, see How to upgrade, downgrade, or cancel.

Pricing

Prices are per organization, billed in USD. Annual is the total for 12 months — there’s no separate “annual discount” line, the headline number is the deal.

Plan	Monthly	Annual
Free	$0	$0
Starter	$19	$190
Pro	$39	$390
Team	$99	$990
Enterprise	Contact us	Contact us

Run, AI, and visual quotas

These reset at the start of each calendar month. Unlimited means no backend cap; fair-use applies on Enterprise.

Limit	Free	Starter	Pro	Team	Enterprise
Cloud test runs / month	300	1,000	5,000	20,000	Unlimited
AI creations / month	100	500	2,000	5,000	Unlimited
AI screenshot checks / month	20	100	500	2,000	Unlimited

What an AI screenshot check is

A visual step compares a screenshot against its baseline in stages. An identical pair and a pair whose pixel difference is under the threshold are decided locally and cost nothing — they are not counted. Only a pair that goes to the AI vision model counts as one check. Past your monthly cap the step fails on the pixel verdict and says the limit was reached, so a spent quota never reads as a passing check. Checks billed to your own provider key (BYOK) are not counted. See Visual regression.

BYOK is a paid feature (Starter and up)

Bring-your-own-key (connect your own OpenAI / Anthropic / Groq / Gemini key) is available on Starter and above, not Free. Even where it’s included, per-user BYOK is gated on the deployment setting BYOK_ENCRYPTION_KEY and off by default — until an operator sets it, the Settings → AI Providers page shows it disabled and the routes return 503. When it’s on, BYOK routes your AI calls through your own provider account. See Using your own OpenAI key and the BYOK concept.

Execution capabilities

Capability	Free	Starter	Pro	Team	Enterprise
Browsers	All 3	All 3	All 3	All 3	All 3
Visual regression	Yes	Yes	Yes	Yes	Yes
Parallel executions	1	1	2	5	Unlimited
Priority run queue	—	—	—	—	Yes

Every plan runs all three engines — Chromium, Firefox and WebKit, Free included (see Browsers). Parallel executions is how many cloud runs your org can have in flight at once; a run over the cap waits in the queue.

Project, test, and data caps

There’s no cap on projects, tests, test sets, folders, or tags at any tier. The only structural caps are inside a single test:

Per-test cap	Free	Starter	Pro	Team	Enterprise
Variables per test	50	50	50	50	50
Data rows per test (parameterized runs)	200	200	200	200	200

These are uniform across plans on purpose — a Free test should look structurally identical to a Team test.

Automation, retention, and seats

Limit	Free	Starter	Pro	Team	Enterprise
Schedules	2	10	Unlimited	Unlimited	Unlimited
API keys	1	1	5	15	Unlimited
Local agents	1	1	3	5	Unlimited
Run history retention	14 days	90 days	180 days	2 years	Unlimited
Seats	1	Unlimited	Unlimited	Unlimited	Custom
Notification contacts	2	10	25	Unlimited	Unlimited

Every plan includes an API key and one local agent, so the CLI, CI integrations, a connected coding agent and local execution all work from Free up. See API keys and Setting up the local agent.

Team and SSO features

Feature	Free	Starter	Pro	Team	Enterprise
Shared projects (across org members)	—	Yes	Yes	Yes	Yes
Org roles (owner / admin / member / viewer)	—	Yes	Yes	Yes	Yes
Google Workspace SSO	—	—	—	Yes	Yes
SAML SSO / audit logs / SLA	—	—	—	—	Talk to us

Shared projects and org roles work on every multi-seat plan (Starter and up — Free has one seat, so there’s nobody to share with). Seats are unlimited on every paid plan; Team adds Google Workspace SSO — see Inviting team members.

Cheat sheet — when to upgrade

• You need more runs, history, or parallelism. Upgrade to a higher tier. (Cross-browser is not a reason: every plan runs all three browsers.)
• You want more local agents, or your own AI key. You need at least Starter (BYOK; Pro 3 agents, Team 5). CI, CLI runs and one local agent work on Free.
• You’re hitting the AI-creation or cloud-run cap. Upgrade to a higher tier.
• You’re running 5+ parallel tests in CI. Pro gives 2× parallel; Team gives 5×.
• You need shared projects or org roles. Starter — they need a second seat, and every paid plan has unlimited seats.
• You need Google Workspace SSO. Team.
• You need SAML, audit logs, or an SLA. Not built yet — contact us.

See also

• How to upgrade, downgrade, or cancel — the actual click-path
• Plan philosophy — why Free is fenced and where each tier earns its price
• Browsers — the engine matrix
• BYOK — moving AI inference to your own provider account