An explicit evidence level
Findings tell detected, reachable and confirmed apart. A reachable flow carries a witness anchored in the code, and a confirmation shows where it comes from: a pentester, a runtime test or DAST.
Tamis vs Semgrep
Semgrep centres on writing rules and an open engine in your CI; its documentation places customer data in the United States. Tamis builds on Opengrep, an open fork of Semgrep, and adds containers, DAST and hosting in the EU or on your premises.
Updated 9 October 2026 · by the Tamis team
Written by the vendor of Tamis, from Semgrep's public documentation (sources below). Method and full comparison.
Why choose Tamis
- Data in the EU or on premises
- Containers and DAST
- An explicit evidence level
Comparison
Reachability scopes differ across vendors (method, languages). “Partly”: open-source engine, proprietary platform.
| Criterion | Tamis | Semgrep |
|---|---|---|
| What is scanned | ||
| Code (SAST) | Yes | Yes |
| Dependencies (SCA) | Yes | Yes |
| Secrets | Yes | Yes |
| Infrastructure as code | Yes | Yes, Terraform |
| Containers | Yes | Not established |
| Deployed app (DAST) | Yes | Not established |
| SBOM export | Yes | Yes |
| License policies | Yes | Yes |
| Prioritisation | ||
| Dependency reachability | Yes, 8 ecosystems | Yes, 11 languages |
| AI fixes | Yes | Yes, beta |
| Import from other tools | Yes, SARIF and 4 formats | Not established |
| Correlation | ||
| Deduplication across scanners | Yes | Not established |
| Code linked to the deployed app (SAST and DAST) | In development | Not established |
| Attack chains across findings | Yes, MITRE ATT&CK | Not established |
| One flaw, every affected project | Yes | Yes, by dependency |
| Hosting | ||
| SaaS in the EU | Yes | No |
| Whole platform on premises | Yes | Not established |
| Scanner on your side, SaaS results | In development | Yes |
| Open-source engine | Partly | Partly |
| Governance | ||
| SSO | Yes | Yes, SAML Enterprise |
| Reports per framework | Yes | Not established |
| Risks and evidence per project | Yes | Not established |
| Beyond scanning | ||
| Threat modelling (DFD, STRIDE) | Yes | Not established |
| Architecture linked to findings | Yes, C4 and DFD | Not established |
| Manual pentest tracking (WSTG, ASVS) and report | Yes | Not established |
| Autonomous AI pentesting | Yes | Not established |
| Retest of a fixed finding | Yes, supported scanners | Not established |
| EBIOS RM workshops | Yes | Not established |
What Tamis does differently
AI fixes that are re-scanned
For supported cases, the proposed fix is re-scanned on a copy of the code to check that the targeted finding is gone. The pull request only opens behind an approval bound to that fix.
Your data, your AI
A local model can give the first triage verdict; the cloud provider is optional and chosen by your administrators. Every suggestion stays a proposal until you decide.
One vulnerability, every affected application
When a shared component is vulnerable, Tamis shows the projects that use it and tracks, project by project, the decision, the fix and its evidence. The same risks and evidence feed your NIS2, DORA or ISO 27001 files.
SaaS at OVHcloud or Scaleway or on premises
SaaS run at OVHcloud or Scaleway in the European Union, or the whole platform on your infrastructure, air-gapped included.
Keep your current scanners
SARIF, Snyk, SonarQube, Nessus and Qualys reports are imported and tracked with the rest, without replacing what already works.
Questions
Does Tamis use Semgrep?
It uses Opengrep, an open-source fork, with its own rules.
Does Semgrep host in the EU?
No: its docs place customer data in the United States.
How can I judge Tamis on my code?
With the free 14-day POC: three projects of your choice, and success criteria agreed together before the first scan.