AI Verify
AI Verify FoundationAI Verify is an AI governance testing framework and toolkit that runs standardized tests on classification and regression models against internationally recognized principles, with plugins that include a process checklist. It runs inside an enterprise's own environment, and its test algorithms can run as separate Python modules. The AI Verify Foundation, a subsidiary of Singapore's IMDA, maintains it.
Tagged self-attested: the organization testing its own model produces the report.
Openness
5 high confidence- license
- Apache-2.0(OSI)
- source
- public
- core features withheld
- no
The whole toolkit, test engine and plugins included, is Apache-2.0, and it is maintained by a foundation under Singapore's IMDA rather than a vendor with a paid tier.
- https://raw.githubusercontent.com/aiverify-foundation/aiverify/main/aiverify-test-engine/LICENSE recorded 2026-09-26
Apache License, Version 2.0, full text, for the test engine
- https://raw.githubusercontent.com/aiverify-foundation/aiverify/main/LICENSE recorded 2026-09-26
Apache License, Version 2.0, full text
- https://raw.githubusercontent.com/aiverify-foundation/aiverify/main/README.md recorded 2026-09-26
"the AI Verify Foundation, a subsidiary under the IMDA"; "AI Verify is an AI governance testing framework and software toolkit"; "a single integrated toolkit that operates within an enterprise environment"
- https://ungh.cc/repos/aiverify-foundation/aiverify/files/main recorded 2026-09-26
Full file listing of aiverify-foundation/aiverify main, 2,348 paths; no ee/, enterprise/, commercial/ or proprietary/ directory
Adoption
1 low confidenceGitHub stars are the only signal. The toolkit's documented install is not a single package, and its test-engine library on PyPI is one component. A star is not a use.
- https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/aiverify-foundation%2Faiverify recorded 2026-09-26
aiverify-foundation/aiverify: 96 stargazers, pushed 2026-03-23
Capability
2 medium confidenceAI Verify runs standardized tests mapped to governance principles, but the organization being assessed runs them and a reader takes the results on trust. Compliance Trestle, a step up, writes evidence to a published standard outside tools can validate.
- https://raw.githubusercontent.com/aiverify-foundation/aiverify/main/README.md recorded 2026-09-26
"validates the performance of AI systems against a set of internationally recognised principles through standardised tests"; "consistent with international AI governance frameworks such as those from European Union, OECD and Singapore"
- https://ungh.cc/repos/aiverify-foundation/aiverify/files/main recorded 2026-09-26
File listing includes third-party-plugins/ccs_plugins/ccs_process_checklist and the stock test plugins
Verified 2026-09-26