Arize Phoenix

An open-source LLM observability platform that unifies traces, evaluations, experiments, and prompt tracking in a single dashboard. Helps you debug agent failures, compare model performance, and optimize costs — all without sending data to a third party.

AI ToolFree Tier: Fully open source; self-hostedCompany: Arize AICategory: Cost OptimizationOpen Source: YesQuick Start: Install with pip install arize-phoenix → Instrument your LLM calls with the Phoenix tracer → Open the local dashboard to view traces, run evals, and track experiments

Arize Phoenix

Open-source LLM observability — traces, evaluations, experiments, prompt tracking in a single dashboard

Visit Arize Phoenix →

An open-source LLM observability platform that unifies traces, evaluations, experiments, and prompt tracking in a single dashboard. Helps you debug agent failures, compare model performance, and optimize costs — all without sending data to a third party.

FREE TIER
Fully open source; self-hosted
COMPANY
Arize AI
CATEGORY
Cost Optimization › Observability
OPEN SOURCE
Yes
PRIVACY
Self-hosted; all telemetry data stays local
TAGS
observabilitytracesevalsdebugging

QUICK START

Install with pip install arize-phoenix → Instrument your LLM calls with the Phoenix tracer → Open the local dashboard to view traces, run evals, and track experiments

BEST FOR

  • ▸LLM teams prioritizing absolute data privacy with self-hosted observability.
  • ▸Engineers needing unified traces, evals, and prompt tracking for LLMs.
  • ▸Developers seeking full control over their LLM observability stack.

NOT FOR

  • ✕Users who prefer managed, cloud-hosted observability solutions.
  • ✕Teams without dedicated DevOps or infrastructure management resources.
Finding similar tools…
← Back to all tools