Overview

The toolkit targets fairness, explainability and robustness, while its framework organizes assessment around 11 principles, including transparency, safety, security, data governance and accountability. It aligns with internationally recognized principles and frameworks from the EU, OECD and Singapore, and maps to NIST AI RMF, its Generative AI Profile, the Hiroshima Process Code of Conduct and ISO/IEC 42001. These give governance teams reference points, but AI Verify does not set ethical standards or certify that a system is safe, unbiased or risk-free.

The AI Verify Foundation is a not-for-profit subsidiary wholly owned by Singapore’s Infocomm Media Development Authority. Founded in 2022, it releases the toolkit under the permissive Apache 2.0 license. That openness makes it a plausible fit for organizations that want to adapt or extend their evaluation workflow rather than buy a closed governance product.

Key features

Testing and workflow

The toolkit provides more than eight technical tests for traditional AI applications. Teams can run them in the Portal or from the command line, upload results, complete process checklists and generate reports. The combination supports self-assessment and third-party testing, and reports can bring developer and compliance-team work together. Standard templates define report layouts, technical tests and process checks; customized reports offer more flexibility when a team’s review process differs.

Data, plugins and Veritas

AIVT 2.0 supports tabular and image data, with inputs supplied as a model file, preprocessing pipeline or model API. That covers several common evaluation setups, but teams working with other data types should not assume the toolkit supports them. Foundation and third-party plugins extend its capabilities. Veritas adds fairness and transparency tests for threshold trade-offs, along with explainability plots; its integration also addresses common safety-baseline and financial-testing requirements from MAS.

Governance and LLM scope

The framework assesses responsible implementation across traditional and generative AI, and supports risk assessments, policy and controls, and compliance-framework work. The Testing Toolkit itself focuses on traditional machine-learning models. Project Moonshot is the separate tool for LLM applications, using benchmark testing and red teaming, including benchmarks recommended in IMDA’s Starter Kit. It offers a Web UI, interactive CLI, library APIs and Web APIs for MLOps integration. Teams choosing AI Verify for governance should treat Moonshot as a separate part of the toolkit family, not as evidence that the traditional testing toolkit evaluates LLM applications.

Pricing

AI Verify is free, with no paid plan or tier distinctions to weigh. Its Apache 2.0 license also permits use and modification, which can suit teams that want to build their own workflow around the toolkit. Free does not mean the toolkit guarantees outcomes: evaluation still depends on the selected tests and the team’s assessment.

Platforms

AI Verify supports API, Linux, self-hosted and web use, with on-premises deployment. Portal and command-line access give teams both a web workflow and a way to run tests from their own environment. The Foundation says electronic storage and transmission of personal data are secured with appropriate security technologies.

Who it's for

Data-science and compliance teams are the clearest fit: the toolkit connects technical tests with process checklists and reporting, while allowing independent work to be combined into a fuller assessment. AI owners, developers, researchers, service providers and companies integrating the toolkit into their systems can also use it. It is less suitable as a standalone answer for teams seeking a safety certification, a guarantee against bias, or direct LLM testing through the Testing Toolkit.

Pros and cons

Open and extensible: Apache 2.0 licensing and third-party plugins give teams room to adapt the toolkit and add capabilities.
Technical and governance work can meet in one report: tests, checklists and contributions from developers and compliance teams support a joined-up review.
Defined data scope: AIVT 2.0 supports tabular and image data, so teams with other data types should verify fit before building a workflow around it.
No safety guarantee: AI Verify explicitly does not define ethical standards or establish that a tested system is free of risk or bias.
LLM testing is separate: teams evaluating LLM applications need Project Moonshot rather than the traditional-model Testing Toolkit.

Alternatives

Compare AI governance tools if you want to survey the category before choosing a framework or testing workflow.

Trusys AI is worth considering for teams seeking a freemium evaluation product with API, self-hosted and web platforms, plus a free Starter plan that allows one application, unlimited functional and security evaluations, 5,000 metrics per month and 30,000 production spans.

Verisum offers a free-forever Explorer plan for teams looking for a freemium alternative with API and web access; its listed maker is based in London.

Prufer may suit teams that want a freemium option with a free trial, but its Free Plan has limited features and quotas, retains audit data for seven days, and provides no uptime guarantee or SLA credits.

OneTrust Third-Party Management is a paid alternative for organizations looking for a product specifically named for third-party management, with pricing based on admin users and third-party inventory.

Deeploy is a paid alternative for teams considering API, self-hosted or web deployment; its Core plan uses a monthly platform fee plus seat-based pricing, has a three-seat minimum and a soft limit of five AI systems per user.

AIGovernr offers a freemium option and a Professional plan at 79.00 USD per month, which includes website scans, findings, gap analysis, action-plan and governance-brief exports, and priority support.

SAS Visual Statistics is a paid alternative for teams seeking enterprise-scale statistical modeling on SAS Viya.

Openlayer Guardrails is a freemium option for teams seeking API, self-hosted or web deployment, with a Basic plan covering one member, five projects, one inference pipeline per project and 20,000 inference logs per month.

Verdict

Choose AI Verify if your data-science and compliance teams want a free, open-source way to run traditional-model tests alongside governance checklists and reporting, with room to extend the workflow. Look elsewhere if you need a guarantee of safety or bias-free results, or expect the Testing Toolkit itself to handle LLM application evaluation.

	AI Verify	ValidMind AI Governance
Free plan	No	No
Free trial	No	No
Paid from	—	—
Open source	No	No
Platforms	api, Linux, self-hosted, Web	api, self-hosted, Web

AI Verify vs ValidMind AI Governance

AI Verify

ValidMind AI Governance