Pricing
Not seats. Not runs.
Pay for living tests.
A maintained test is one we keep alive, healed and reporting. That is the only number on your invoice — and your QA lead already knows it by heart.
Starter
€149/mo
A living smoke suite for one product.
- 25 maintained tests
- 1 project
- Daily schedule + on-demand runs
- Fair use 1,500 runs/mo
- Shared agent pool
- Self-healing included
- Email support
Growth
€490/mo
Full regression coverage that runs on every commit.
- 100 maintained tests
- 5 projects
- On-commit (smart test selection) + schedules
- Fair use 25,000 runs/mo
- Shared pool, priority queue
- Self-healing + healing diff review
- Slack support
Scale
€1,490/mo
QA as infrastructure, with guarantees.
- 400 maintained tests
- Unlimited projects
- Unlimited runs (fair use)
- 4 reserved parallel agent slots
- 99.5% run-SLA
- SSO + audit export
- Self-healing + diff review
Dedicated Fleet
from €1,980/mo (2+ agents)
Your own named SparkNodes — hosted or on-prem.
- Unlimited maintained tests
- Unlimited projects & runs
- Named agents, hosted or on-prem
- VPC / site-to-site connectivity
- Data residency guarantees
- Custom local models
- Self-healing included
Free dev tier: 3 maintained tests, 300 runs/mo, forever — no card. Wire us into CI before you ever pay.
Overage: €6 / €5 / €3.50 per extra test per month on Starter / Growth / Scale — always priced so the next plan up is the better deal.
Annual: two months free (−17%) on every plan.
Trial: 14 days, 5 maintained tests, shared fleet. No card required.
The manifesto
Why we price per test — and give runs away.
Every pricing model teaches customers a behavior. Per-run pricing teaches you to test less. Ours teaches you to cover everything.
It scales with value, not with our costs
You buy coverage — tests that are alive, healed and reporting. Runs are ~free on hardware we own, so we price the thing you value and give the runs away.
It is legible to buyers
A QA lead knows exactly how many tests they have. Finance can forecast to the euro. No seat audits, no per-minute meters, no surprise invoice after a busy release week.
It aligns incentives
Vendors that meter runs teach you to run less. We only win by keeping tests alive — and every extra run you schedule makes our healing smarter. So run on every commit.
Compare
Every plan, side by side.
| Plan | Starter | GrowthMost popular | Scale | Dedicated Fleet |
|---|---|---|---|---|
| Monthly price | €149 | €490 | €1,490 | from €1,980 (2 agents) |
| Maintained tests | 25 | 100 | 400 | Unlimited |
| Extra tests (overage) | €6 /test /mo | €5 /test /mo | €3.50 /test /mo | — |
| Projects | 1 | 5 | Unlimited | Unlimited |
| Run triggers | Daily schedule + on-demand | On-commit + schedules | Unlimited (fair use) | Unlimited |
| Fair-use runs | 1,500 /mo | 25,000 /mo | Unlimited (fair use) | Unlimited |
| Agents | Shared pool | Shared pool · priority queue | 4 reserved parallel slots | Named SparkNodes · hosted or on-prem |
| Self-healing | Included | Included + diff review | Included + diff review | Included + diff review |
| Support | Slack | Slack | Custom | |
| Run SLA | — | — | 99.5% | Custom |
| SSO + audit export | — | — | Included | Included |
| VPC / site-to-site | — | — | — | Included |
| Data residency | — | — | — | Included |
| Custom local models | — | — | — | Included |
| Trial | 14 days · 5 tests · no card | 14 days · 5 tests · no card | 14 days · 5 tests · no card | Guided pilot |
FAQ
Fair questions, straight answers.
What counts as a maintained test?
A test case SparkTest keeps alive: codified into a Puppeteer script (up to 25 steps / 5 minutes of runtime), executed on your schedule, healed when it breaks, and reporting into your console. Archived or quarantined tests don't count against your plan; longer journeys count as multiple tests.
What happens when a test breaks?
The agent feeds the failure and DOM context to a local LLM, patches the script and re-runs, with a bounded number of attempts. The run is reported as healed, with the full script diff for your review. Healing is included in every plan — you never pay to fix a broken test.
Where do agents run?
On SparkTest's fleet of SparkNodes — compact, purpose-built AI appliances in our own racks, each running a local LLM. Agents connect outbound over HTTPS only — no inbound ports, no VPN. On the Dedicated Fleet plan, your named agents can be hosted by us or installed on-premises.
Is my data safe?
Codify and heal run on-device, on hardware we own — your DOM, flows and failures never reach a third-party model API. Scripts, logs and screenshots live in the control plane on the SparkTech Backbone platform; Scale and Dedicated add SSO, audit export and data-residency options.
Can I export my scripts?
Yes. Every test is a plain, versioned Puppeteer script — readable and auditable. Export them anytime; they run anywhere Node runs. If you leave, your test suite leaves with you.
What sites can be tested?
Anything a real browser can reach: production sites, staging environments behind basic auth or IP allowlists, and — with Dedicated Fleet — apps on private networks via VPC or site-to-site connectivity.
Coverage you can forecast. Tests that stay green.
14-day trial · 5 maintained tests · shared fleet · no card required.