Phoenix
Arize's source-available tool for tracing, evaluating, and debugging LLM and agent apps.
Phoenix, from Arize, is an observability and evaluation tool built on OpenTelemetry. It traces LLM and agent applications, runs evaluations, and is strong at the debugging end: inspecting and comparing runs to find where quality broke.
It self-hosts and runs locally, and pairs with Arize's larger production-monitoring platform.
Where it fits
A fit when you want open, OpenTelemetry-based tracing and evaluation with a focus on debugging agent behaviour.
Where it does not fit
For long-term production monitoring at scale, the hosted Arize platform is the heavier sibling to grow into.
Alternatives to Phoenix
The closest options on the map, and where each fits instead.
-
Langfuse LLM observability
A strong default whenever an AI system reaches production and you need to see what it actually did, call by call.
Unlike Phoenix Open sourceHosted APIMaturity: established
-
Opik LLM observability
A fit when you want an open-source, self-hostable platform spanning tracing, evals, and production monitoring.
Unlike Phoenix Open sourceHosted API
-
OpenLLMetry Observability standard
A fit when you want LLM tracing as standard OpenTelemetry data, feeding an observability backend you already use.
Unlike Phoenix Open source
-
Helicone LLM observability
A fit when you want fast, low-effort visibility into model calls, costs, and latency through a simple proxy.
Unlike Phoenix Open sourceHosted APIMaturity: established
Pairs with
Commonly used alongside it.
