Nothing to scope, nothing to schedule
A target and the depth you want is the whole intake. There is no proposal to sign off and no slot to wait for, so the run starts when you start it rather than when a calendar opens up.
pentest as a service
A pentest you subscribe to, not one you book.
Testing you take out rather than book: always available, running continuously instead of once a year, and the report is never the thing you are waiting on. Point it at a target, pick the depth, and findings come back the same day.
01 / The service
What a subscription replaces.
Most of what a pentest costs is not the testing. It is the scoping, the waiting, the document, and the work of turning that document into tickets.
One subscription, every application
Add targets as you build them. A new service does not turn into a new engagement with its own budget line and its own three weeks of lead time.
Findings during the run, not after it
They arrive as they are proven, so the first one is on someone's desk while the rest is still going. Nothing sits in a draft waiting to be collected into a deliverable.
The report is an output, not a project
Every run ends with the report an engagement ends with, generated from what actually happened. It is the document teams hand an auditor for ISO 27001 or SOC 2, and nobody writes it.
Remediation is part of the service
With your repository connected the patch is written against your code and opened as a pull request. Fixing is not a second project that starts once the testing is over.
You pay for what it proves
Start free. Past that the bill follows findings that carry evidence, not days booked or scans started.
02 / Getting started
From a URL to a running service.
Each step gives the agent more to work with. None of them is a prerequisite for the one before it.
01
Name the target
A URL is enough for a black box run. The agent maps what is reachable and works from the position an outside attacker starts from.
02
Hand over as much as you want
Test accounts make it grey box, which is where broken access control surfaces. Your repository makes it white box, with the code path behind an endpoint read rather than inferred.
03
Send findings where you work
Jira, Linear, GitHub Issues, Slack, Teams or Discord. Nobody has to learn another dashboard before they can pick the work up.
03 / Cadence
Continuous by default, at the interval you pick.
This is the part a booked engagement cannot do. The depth you chose is captured with the schedule, so runs stay comparable and the tenth goes as deep as the first.
01
On every push
The run triggers from the repository integration, so the change that introduced a problem and the finding that proves it arrive together rather than a quarter apart.
02
Daily
For targets that move faster than their release notes. Overnight runs mean the morning starts with what is true now, not with what was true at the last audit.
03
Weekly
The cadence most teams settle on. Fast enough that a regression is caught inside the sprint it landed in, quiet enough that findings stay readable.
04
Monthly or quarterly
Where a scheme asks for a documented interval, the schedule is the document. Each run ends with its own dated report, so the interval is evidenced rather than asserted.
04 / What comes back
What every run leaves behind.
Evidence on every finding
The request and the response where it exploited something, the file and the line where it read something. Judge it yourself rather than trusting a severity score.
A pentest report, dated
Generated when the run ends, not typed up in the week after. That is what makes the interval provable: a report per run rather than one a year.
A pull request per fix
Written with the safe functions your codebase already uses, opened on its own branch, and yours to review, trim or reject.
One finding, tracked across runs
Something still open on the next run stays the same finding rather than arriving again as a duplicate, so how long it has been live is visible instead of counted by hand.
A record of what was tested
Every run is kept with its transcript and its frozen configuration. A finding that gets challenged is settled by replaying it rather than by arguing about tooling.
05 / Questions
Questions, answered.
ready when you are
Put a target under test today.
A URL is enough to start. Add logins and it goes deeper, connect the repository and the fix comes back with the finding.