Skip to content
ScanPry

Early access, by invitation

Automated security testing for your websites and source code

ScanPry runs proven open-source scanners against the websites you prove you control and the code you upload, then turns their results into one list of findings, each with the evidence behind it.

Accounts are by invitation while we are in early access.

One place for what every scanner found

Each engine reports in its own format and its own words. ScanPry reads them all into one model, so you work from a single list.

  • One list, not a report per tool

    Every engine's results are validated and normalized into one finding model, then deduplicated across engines and across scans.

  • Evidence you can check

    Each finding keeps what shows it: the request and the response, or the file and the line. A secret found in your code is stored redacted, never in full.

  • Severity and weakness class

    Findings carry a severity and are grouped by CWE weakness class, so the overview shows what kind of problems you have, not only how many.

  • Change over time

    A finding a later scan no longer sees is marked fixed, and one that returns is marked regressed. Accepting a risk or calling a result a false positive records who decided, and why.

  • Reports for people and for tools

    HTML, PDF and Excel reports in English or Vietnamese, and JSON, SARIF and CSV exports for your own pipeline.

  • Built for teams

    Organizations, workspaces and projects keep clients and products apart, with roles at each level. Two-factor authentication is required of anyone who can start a scan.

How a scan works

  1. Add a website and prove you control it

    Publish a value we issue in a DNS TXT record, a file or a meta tag. Nothing scans a website until the proof checks out, and every scan checks it again when it starts.

  2. Choose what to run

    Quick, Standard or Deep for a website. For code, upload your source as a zip archive. Each workspace decides which engines it allows.

  3. Engines run in isolation

    Each engine runs in its own short-lived container, with no access to our database or internal network, and the container is destroyed when it finishes.

  4. Review, fix, scan again

    Findings arrive in one list with their evidence. Scan again to confirm a fix, and export a report when someone needs one.

Proven open-source engines, pinned and recorded

We run the scanners security teams already rely on. The exact version of every engine is pinned and recorded with each finding, so a result can be traced to what produced it.

Websites

  • ffuf
  • Nuclei
  • OWASP ZAP
  • our own HTTP checks

Source code

  • ast-grep
  • Bandit
  • Checkov
  • eslint-plugin-security
  • Gitleaks
  • gosec
  • Kingfisher
  • Psalm
  • Trivy

Safe to point at the internet

A platform that sends requests to websites must not become a way to attack someone else's. These rules are enforced in the code, not left to policy.

  • Only a website you have proven you control can be scanned, and the proof is checked again when each scan starts.
  • A target that resolves to a private, internal or cloud-metadata address is refused before anything is scanned.
  • Uploaded code is read and never run: no build, no install, no scripts.
  • Scanner containers hold no database credentials and cannot reach our internal network.
  • Who started each scan, against which website and under which proof, is kept in an audit trail that application code cannot rewrite.

See what your websites and code expose

Sign in if you have an account. If not, write to us and tell us what you would like to test.