self-hosted vs hosted

Self-host it free. Or skip the bills.

The engine is open source and fully capable — run it yourself whenever you want. The hosted cloud exists for one reason: predictable, hands-off scanning. No API keys, no surprise LLM token bills, and you only pay per live host you actually scan.

What does it really cost?

Self-hosting is free to run, but you pay your LLM provider for every token the agent burns — and agentic, multi-phase scans are token-hungry and hard to predict. Move the sliders to your setup.

Self-hosted
$600/mo

LLM tokens only. Excludes your servers, maintenance, and time.

Hosted
$14.00/mo

20 credits · pay-as-you-go. From $8.17/mo on a subscription. All-in.

At this volume, hosted is about $586/mo cheaper — and that ignores the servers, upgrades, and rate-limit babysitting you skip. No API keys, no bill surprises.

Estimates only, for comparison. Token usage varies widely with model, scope, and scan depth — tune the sliders to your own numbers. Hosted figures use current published pricing (pay-as-you-go at $0.70/credit, subscriptions from $0.41/credit).

Side by side

Both run the exact same 22-phase, exploit-verified engine. The difference is everything around the scan.

 Self-hosted (OSS)Hosted cloud
Price to startFree, open source (Apache-2.0)One full scan from $1 · credits never expire
LLM API keyBring & manage your ownIncluded — no keys to wrangle
Cost per scanRaw LLM tokens — variable, can spike on deep/agentic runs1 credit per live host — predictable, billed on results
Setup & opsYou install, update, and run the engine + toolchainNothing to run — scan in ~60 seconds
Out-of-band infra (SSRF, blind RCE, XXE)You stand up your own OOB/collaborator serverManaged OOB included
Scale & concurrencyOne box you nurse; you handle rate limitsManaged queue, multiple backends, auto-scaled
Scheduled & recurring scansWire up your own cronBuilt in (daily / hourly)
Team, RBAC & shared creditsSingle-userWorkspaces, roles, shared credit pool
UpdatesPull & rebuild yourselfAlways on the latest engine
Data residency & privacyEverything stays on your infrastructureRuns on our infrastructure (DPA available)
Air-gapped / offline useSupported (local models via Ollama)Not applicable — it's a cloud service
SupportCommunity / GitHub issuesEmail support with SLAs on paid plans

★ = where self-hosting is the better fit. If data must never leave your network, or you want to run fully offline, self-host — that's exactly what it's for.

Try a full scan for $1

No subscription, no API keys, credits never expire. See exactly what an exploit-verified report looks like — then decide.