About Fiddler AI
Fiddler AI sells the AI Control Plane, a platform for first-party and third-party agents from the creation layer to production. Fiddler combines continuous evaluation, monitoring, enforceable policy and auditable governance for enterprises. Fiddler Centor Models power guardrails and evaluations in-environment, with no external LLM API calls.[2][3]
Fiddler Agentic Observability monitors every layer of the agentic hierarchy from application health to agent interactions to individual spans. Fiddler ML Observability provides root cause analysis and fairness metrics such as disparate impact, demographic parity and equal opportunity. Fiddler Guardrails enforce policy inline, at the agent's request and response path, in under 80ms. Fiddler detects and redacts PII, PHI and secrets on the same request, before a prompt reaches the model and before a response reaches the developer. The Free plan includes real-time guardrails powered by Fiddler Centor Models at under 80ms latency, protecting against hallucinations, toxicity, PII/PHI, prompt injection and jailbreak attempts. Fiddler tracks 100+ out-of-the-box and custom metrics including hallucination, toxicity, PII/PHI leakage and jailbreak attempts.[2][3][4][5][6]
Fiddler AI provides centralized governance and oversight for agents and predictive models across the enterprise, with a registry of your agents, enforced guardrails and the audit evidence needed to meet evolving regulatory requirements. Agent Registry lists every agent application in Fiddler, including first-party, third-party and coding agents. Fiddler records evaluation and policy outcomes with full traceability and generates audit evidence aligned with frameworks and regulations such as GDPR, HIPAA, NAIC, SR 11-7, NIST AI RMF, ISO/IEC 42001 and the EU AI Act. Guardrails run in your own environment, so prompts and responses never leave your infrastructure. Teams can monitor predictive models and agents for drift, bias and performance degradation, run deep diagnostics to find root causes, and create custom reports for MRM and compliance reviews. Fiddler extends governance to the creation layer, monitoring the IDE, CLI and MCP boundary to validate tool invocations against allowlists, scan for secrets, and prevent coding agents from committing insecure code or accessing unauthorized repositories.[5][7]
Fiddler is offered in SaaS, VPC and AWS GovCloud. Fiddler can be deployed in any secure cloud environment or air-gapped. Fiddler states it is SOC 2 Type II compliant, with role-based access control and SSO. Fiddler states it is HIPAA ready. Fiddler AI Observability can run within Amazon SageMaker Studio. Fiddler is model- and cloud-agnostic, working across Azure OpenAI, Amazon Bedrock, LangGraph and 100+ other providers. The Developer plan is usage-based at $0.002 per trace and adds unified observability for agentic and predictive systems, custom evaluators and SSO. The Enterprise plan adds flexible deployment of SaaS, VPC or on-premises, along with white-glove support with a named Customer Success Manager and customized onboarding.[3][5][8]
Fiddler AI is used by organizations deploying agents across all industries, as well as highly regulated industries like financial services, healthcare, insurance and government. The typical users are software and AI engineers, data scientists, platform and DevOps teams, and security and compliance teams. Fiddler positions its AI Observability and Security platform as the Unified AI Command Center for insurers. Fiddler provides a unified AI observability and security platform for healthcare organizations. Fiddler describes itself as the control plane for AI agents in financial services, covering monitoring, policy and governance. Financial institutions deploying AI across underwriting, fraud and trading rely on Fiddler for unified observability, policy enforcement and regulatory-ready audit trails. Fiddler monitors Natural Language Processing and Computer Vision models to track and get real-time alerts on anomalies and data drift. Fiddler explains predictions in ML models powering autonomous vehicles, such as aerial drones and unmanned underwater vehicles. Fortune 100 organizations use Fiddler for agentic and predictive applications. For customer-facing agents, Fiddler governs agents built in-house or integrated from third-party providers in a unified dashboard with audit evidence for GDPR, CCPA, HIPAA and NAIC compliance.[2][3][4][5][6][8][9][10]
Fiddler's founders began with explainability and then expanded to observability to track how models change over time. Fiddler describes itself today as the Control Plane for AI Agents. IQT first invested in Fiddler in 2019.[2][4][6][8][10][11]
Fiddler AI on the record