Giskard is a Paris startup whose open-source and enterprise testing tools find hallucinations, security holes and bias in AI models and LLM applications.
As companies ship LLM features, they need tests that go beyond unit tests: does the model hallucinate, leak data or fail on edge cases? Giskard’s open-source library generates automated tests and scans for vulnerabilities, while its enterprise Hub adds continuous evaluation, monitoring and compliance workflows for production AI. The company is active in the ML testing research community and its tooling integrates with standard CI pipelines.
The open-source testing libraries are free; Giskard Hub is a commercial product with team and enterprise plans based on scale and support.
ML engineers, QA teams and AI governance functions that need verifiable evidence that their models and LLM apps behave safely.
Giskard is the safety net for teams shipping models and LLM features into production.
Install the open-source library and run your first scan in minutes; adopt the Hub when you need shared workflows and compliance reporting.
Teams catch quality and safety regressions before customers do, and produce audit-ready evidence for governance processes.
Open-source tools install in minutes and run in CI; the Hub adds team workflows and compliance reporting for enterprise programs.
Automated tests surface risks but don’t fix them, and coverage depends on thoughtful test design.
Teams appreciate catching hallucination and bias regressions before customers do.
Shipping AI without testing is how brands end up in headlines. Giskard makes systematic testing practical: run it in CI, scan for the failures that matter and keep evidence for governance. It won’t fix what it finds, but knowing is most of the battle.
Yes for the open-source libraries; the Hub platform for continuous testing and monitoring is subscription-based.
Yes — tests and scans run programmatically and integrate with CI pipelines, and the Hub exposes APIs for evaluation workflows.
Against manual red-teaming or generic observability, Giskard automates systematic testing of models and LLM apps with compliance-ready reporting.
Diffblue Cover uses reinforcement-learning AI to write complete, runnable JUnit tests for Java codebases automatically. OverviewDiffblue Cover automates one of the least-loved…
SonarQube is the industry-standard continuous code quality and security platform, analyzing code in 30+ languages across the development pipeline. OverviewSonarQube, from SonarSource,…
Metabob combines graph neural networks with LLMs to detect and explain bugs in code during review, catching issues static analyzers miss. OverviewMetabob's…
