One pool for everything hosted: compute, crawling, monitoring and residential proxies, spent however your work breaks down. On every paid plan the pool is bigger than the price, and AI gets its own monthly allowance on top. Record and replay on your own machine for $0, unmetered, on every plan including Free.
Every plan includes unlimited local runs. Each plan adds a monthly pool of cloud usage for hosted runs, crawling, monitoring and residential proxies, worth more than the plan costs, plus a separate monthly allowance of managed AI. No per-unit math.
Record a workflow once, then replay it on your own machine as often as you want. Local runs never touch your included usage, are never metered, and never reach a bill. That holds on Free exactly as it does on Scale.
Build and run website APIs on your own machine.
Keep your first APIs and schedules online.
Operate client-grade authenticated APIs.
Run a production team and your own agent fleet.
High-volume APIs across a large agent fleet.
The usage pool funds cloud compute, crawling, monitoring and residential bandwidth; managed AI has its own monthly token allowance, and the pool never pays for AI. Local and BYO-agent runs stay free and unlimited on every tier. Need more? Add funds any time, at the same rates; they never expire.
Same crawl, extraction, and automation, with residential included and local workflow runs free.
Real residential egress is funded from your included usage on every paid plan. Rivals meter proxy bandwidth from the very first byte.
Record once, replay on your own machine at $0, unmetered on every plan, Free included. You bring the hardware; we never charge for it.
Everything in the cloud except AI draws from a single monthly pool, priced in plain dollars at the rates below. Past the pool, the same rates apply to your prepaid balance — there is no price change at the edge, and nothing is charged automatically.
$0.12 / browser-hour of hosted run time, and at least $0.001 per cloud run, which covers its first 30 seconds. Runs on your own machine cost $0 and are never metered.
$10 / GB and included in every paid plan. On Free, residential draws from your balance only.
from $3.20 / million tokens past your plan’s own monthly AI allowance. Larger models cost more, and the price is shown before you run. The usage pool never pays for AI.
That usage is covered by the plan — you pay the plan price and nothing on top, and only usage beyond the plan would come out of a prepaid balance. Figured at about $0.04 for a light cloud run; no cloud run costs less than $0.001. Heavier runs, crawling, monitor checks and residential bandwidth draw on the same pool, so a real month lands near this rather than exactly on it. Managed AI is counted separately, against its own allowance. Local runs stay at $0 whatever the volume.
Cloud spend cap, on by default. Cap how much the cloud can spend per period. At the cap, cloud runs pause rather than overrun. Local runs keep going, free. We refund genuine surprise charges. See billing details →
Exact numbers, read from the live plan ladder. Everything hosted except AI draws from one pool.
| Feature | Free | Starter | Pro | Growth | Scale |
|---|---|---|---|---|---|
| Usage & AI | |||||
| Included usage / month | $5 | $20 | $65 | $260 | $650 |
| AI tokens / month (own allowance) | 200K | 1M | 3M | 10M | 20M |
| Billed yearly (2 months free) | $150 / yr | $490 / yr | $1,990 / yr | $4,990 / yr | |
| Unused usage rolls over (1 month) | Included capacity | Included capacity | Included capacity | Included capacity | |
| Residential proxies | Pay-as-you-go | Included | Included | Included | Included |
| Add funds (never expire) | Included capacity | Included capacity | Included capacity | Included capacity | Included capacity |
| Cloud spend cap (on by default) | Included capacity | Included capacity | Included capacity | Included capacity | Included capacity |
| Cloud capacity | |||||
| Concurrent browsers | 3 | 5 | 10 | 25 | 75 |
| Live browsers at once | 1 | 2 | 3 | 5 | 10 |
| Connected agents (you operate) | 1 | 1 | 2 | 5 | 20 |
| CAPTCHA + managed 2FA solving | Included capacity | Included capacity | Included capacity | ||
| Queue priority (busy fleet) | Standard | Priority | High | Higher | Top |
| Runs start within | 120 s | 60 s | 30 s | 20 s | 10 s |
| Monitoring | |||||
| Static-page monitors | 25 | 100 | 500 | 2,000 | Unlimited |
| Browser (JS) monitors | 5 | 20 | 100 | 400 | Unlimited |
| Monitoring load (checks / minute) | 10 | 40 | 75 | 200 | 600 |
| Fastest check · static page | 5 min | 1 min | 1 min | 30 s | 10 s |
| Fastest check · browser (JS) | 15 min | 10 min | 10 min | 5 min | 2 min |
| Local: always free | |||||
| Local / self-hosted runs | Unlimited | Unlimited | Unlimited | Unlimited | Unlimited |
| Local compute charge | $0 | $0 | $0 | $0 | $0 |
| Intranet / localhost / VPN reach | Included capacity | Included capacity | Included capacity | Included capacity | Included capacity |
| Build surface | |||||
| Workflows (record / AI / discover) | Included capacity | Included capacity | Included capacity | Included capacity | Included capacity |
| Cloud recording sessions / month | 10 | 50 | Unlimited | Unlimited | Unlimited |
| 30+ step types | Included capacity | Included capacity | Included capacity | Included capacity | Included capacity |
| REST endpoints + MCP tools | Included capacity | Included capacity | Included capacity | Included capacity | Included capacity |
| Automations & webhooks (HMAC) | Included capacity | Included capacity | Included capacity | Included capacity | Included capacity |
| Personas (2FA: TOTP + mailbox OTP) | Included capacity | Included capacity | Included capacity | Included capacity | Included capacity |
| Personas (saved sign-ins) | 3 | 10 | 25 | 100 | Unlimited |
| Secrets vault + external managers | Included capacity | Included capacity | Included capacity | Included capacity | Included capacity |
| Marketplace templates | Included capacity | Included capacity | Included capacity | Included capacity | Included capacity |
| Sell on the Marketplace | Included capacity | Included capacity | Included capacity | ||
| Team & limits | |||||
| Team members | 1 | 1 | 3 | 10 | Unlimited |
| Managed storage | 1 GB | 2 GB | 5 GB | 50 GB | 200 GB |
| Scheduled workflows | 2 | 10 | 25 | 100 | Unlimited |
| Run history (with extracted data) | 90 days | 90 days | 180 days | 365 days | 730 days |
| API rate limit | Standard | Standard | Higher | Higher | Highest |
| Published endpoints & MCP tools | 3 + 3 | 10 + 10 | 25 + 25 | 100 + 100 | Unlimited |
| Relay fair-use (calls/day) | 2,000 | 10,000 | 25,000 | 50,000 | 100,000 |
Local / BYO runs are free and unlimited on every tier: $0, never metered. Monitoring load is sustained checks per minute across all your monitors; a browser (JS) check counts as 5, and every check is paid from the usage pool. Relay fair-use is a daily ceiling on calls relayed to your published endpoints — sized so honest integrations never notice it, with a heads-up at 80%; past it, calls pause until the day resets (UTC). Enterprise adds SSO, audit logs, an SLA, unlimited published endpoints & MCP tools, and a raised fair-use ceiling.
Writ runs on your own accounts, with your own credentials and data, on sites you are authorized to use.
Free · $0
Metered
Faster monitor cadence is a paid capability: the quickest check intervals live on higher plans.
Two ways forward: a managed contract, or your own infrastructure.
Custom usage volume and committed-use pricing, monitors down to 10-second cadence, SSO and audit logs, a contractual SLA, and a DPA.
Run the whole stack on your own infrastructure. Every run is local: $0, unmetered, no cloud bill at all. Open-source core.