Fiddler AI

Fiddler AI monitors and evaluates AI agents, LLM applications and traditional machine learning models in production, and enforces policies with real-time guardrails. Its buyers are enterprise AI and ML teams in government, finance, healthcare and insurance.

The platform traces each agent decision, runs continuous evaluations from testing to production, and offers model drift and performance monitoring for predictive models. Fiddler's own Centor models handle evaluations and guardrail checks quickly and at low cost, and governance features track risk and compliance.

A Developer tier is billed per trace and adds custom evaluators, bring-your-own judge, SSO and role-based access. Enterprise contracts add hardened guardrails, VPC or on-premises deployment and a named success manager.