Enterprise AI observability, security, and governance control plane for models and agents
EU AI ActNIST AI RMFGDPR
Arthur
AI performance, evaluation, and governance platform for ML, generative, and agentic systems
What Arthur does
Arthur (branded Arthur AI) is a New York-based AI performance and governance company, founded in 2018, that helps enterprises launch, secure, and optimize AI systems. Its platform provides continuous evaluation across the AI lifecycle - pre-production testing, runtime guardrails, and always-on production monitoring - spanning traditional machine learning, generative AI, and increasingly agentic AI. For classical models Arthur tracks drift, accuracy, precision/recall, and bias; for LLMs and agents it measures hallucination and groundedness, enforces acceptable-use and data-security policies, detects prompt injection and PII exposure, and visualizes agent traces and tool selection. In December 2025 Arthur added Agent Discovery & Governance for cataloging and policy-enforcing autonomous agents in production. The company open-sourced the Arthur Engine, a real-time evaluation engine, lowering adoption friction for teams that want self-hosted guardrails. Deployment options include multi-tenant SaaS, single-tenant VPC/on-prem, and hybrid data-plane/control-plane architectures, with SOC 2 Type II, HIPAA BAAs, RBAC, and SSO for regulated buyers. Arthur is best suited to data-science, risk, and compliance teams in banking, healthcare, and insurance.
Key capabilities
- Pre-production, runtime, and production evaluations
- Real-time guardrails (prompt injection, PII, toxicity)
- Hallucination and groundedness detection
- Agent trace visualization and tool-selection metrics
- Drift and performance monitoring for classical ML
- Open-source Arthur Engine
- Agent discovery and governance
Best for
Enterprises operationalizing generative and agentic AI that want flexible deployment (SaaS, VPC, on-prem) and an open-source evaluation engine alongside governance controls.
Limitations
Named enterprise references are limited publicly; funding has not advanced past its 2022 Series B, and the rapid pivot toward agentic AI means several governance features are relatively new.
Framework coverage
| Framework | Type | Supported |
|---|---|---|
| EU AI Act | Regulation | Yes |
| NIST AI RMF | Voluntary framework | Yes |
| SOC 2 | Control framework | Yes |
| HIPAA | Regulation | Yes |
Compare Arthur
Head-to-head against the closest tools in its category.
Arthur alternatives
Other tools solving a similar problem in Observability & Monitoring.
AI observability and evaluation platform for ML models, LLM apps, and agents
SOC 2HIPAAGDPROpen-source
AI control platform combining ML observability with real-time guardrails for GenAI
SOC 2GDPRHIPAA
AI model testing roots now applied to document workflow automation for regulated industries
SOC 2HIPAA
Open-source-led testing, evaluation and monitoring for ML models and LLM applications
SOC 2GDPRHIPAAOpen-source
AI quality, testing, and monitoring platform for evaluating and safeguarding models in production
ISO/IEC 42001GDPRHIPAAOpen-source
AI Governance Tool Selection Kit
A vendor-comparison worksheet plus EU AI Act, NIST AI RMF and ISO/IEC 42001 requirement checklists — so you can shortlist tools against the obligations that actually apply to you.
Free. No spam — unsubscribe anytime.