Arize AI

Arize AI: LLM Observability and AI Model Monitoring

Arize AI is the leading platform for LLM observability and AI model monitoring, built on OpenTelemetry standards. As AI agents become production infrastructure, Arize provides the observability layer that engineering teams need to understand, debug, and improve AI system behavior at scale.

Phoenix: Open-Source LLM Observability

Arize's Phoenix (9,100+ GitHub stars) is the open-source foundation for AI observability during development. It traces every step of agent execution — prompts sent, tools called, memory accessed, routing decisions, and LLM outputs — giving developers unprecedented visibility into agent behavior.

Key Features

• Phoenix OSS: Open-source LLM tracing and evaluation for development
• Arize AX: Enterprise production monitoring at scale
• Agent visualization: Trace every step of AI agent execution
• LLM evaluation: Built-in evals for hallucination, relevance, toxicity
• OpenTelemetry: Standards-based instrumentation for portability
• 1T+ spans/month: Production-scale processing capability
• Multi-model support: OpenAI, Anthropic, Google, open-source models

Pricing

Phoenix is free and open source. Arize AX enterprise pricing available on request.