The open web, at concurrency.
Your own machine runs a workflow perfectly well — until you need forty of them at once, at three in the morning, against a site that would rather not be automated. That is what the fleet is for.
Four things a laptop cannot do.
Everything else is identical — the same workflow, the same steps, the same output. The cloud sells reach and scale, never features you would otherwise be denied.
Concurrency, cadence, and the tougher sites.
These are the three levers that actually differ between tiers. Everything below is the number the app enforces, not a marketing rounding.
| Plan | Concurrent browsers | Fastest monitor | Longest run | Longest recording | Challenges + residential |
|---|---|---|---|---|---|
| Open Source | 3 | 5 min | 2 min | 10 min | — |
| Starter | 5 | 1 min | 4 min | 15 min | — |
| Pro | 10 | 1 min | 5 min | 30 min | Included |
| Growth | 25 | 30 s | 10 min | 45 min | Included |
| Scale | 75 | 10 s | 15 min | 60 min | Included |
| Enterprise | 150 | 10 s | 15 min | 60 min | Included |
The free tier runs three browsers at once and checks every five minutes — enough to prove the thing works before you pay for anything.
One pool, published rates.
Every plan includes a monthly pool of included usage, larger than the price on every paid plan. The same pool funds running time, crawling, monitoring and residential bandwidth, so you spend it on whatever the month actually needs. Managed AI has its own monthly token allowance and never draws the pool. Past the pool you top up and pay exactly the same rates.
A run on your own machine stays free on every plan, including the free tier — even when the cloud is what dispatched it.
A run that signs in or touches credentials executes on the isolated pool — a fresh sandboxed browser, destroyed after the run — at 5 credits a minute instead of 2. The tier is decided from where the run actually executed, never from a flag a client could set.
Where your runs execute, on the record.
We would rather point you at the register than paint a map. The hosting providers we engage, what each one touches and the region it sits in are listed in the sub-processor register, and that list is what our DPA binds us to.
See the sub-processor register
Your workflows, credentials and results belong to your account. We do not sell data and we do not train models on your runs.
Questions people actually ask
Do I have to use the cloud?
What is a credit worth?
What happens when the pool runs out?
Which region do runs execute in?
Is a failed run charged?
Let the fleet take the awkward runs.
Start free. Your own machine stays unmetered whatever you choose.