Arize Phoenix

Arize Phoenix is an open-source AI observability and evaluation platform for developers building LLM applications and agents. It helps teams trace AI workflows, score outputs, evaluate regressions, iterate on prompts, and compare experiments using real application data. The platform is useful for SaaS founders and technical teams building chatbots, agents, RAG systems, internal copilots, research tools, and support automation. Instead of guessing whether a prompt or model change improved the product, Phoenix gives teams a workflow for debugging and quality measurement. Phoenix supports tracing, evaluations, prompt management, datasets, experiments, and integrations with popular AI frameworks and providers. It can be used locally, self-hosted, or through Phoenix Cloud depending on a team's workflow and data needs. For FutureStack, Arize Phoenix is a strong Developer Tools listing because observability and evals are essential for production AI apps. It is best for builders who care about understanding what happened inside an AI workflow and improving quality over time with evidence.