Loading tool details…
Arize Phoenix is an open-source AI observability and evaluation platform for tracing, testing, and improving LLM applications and agents.
Arize Phoenix is an open-source AI observability and evaluation platform for tracing, testing, and improving LLM applications and agents.
Arize Phoenix is an open-source AI observability and evaluation platform for tracing, testing, and improving LLM applications and agents. It uses OpenTelemetry-compatible tracing, supports datasets and experiments, runs evaluations, and includes a Prompt IDE for examining and iterating on prompts. The key distinction is deployment flexibility: a developer can run Phoenix locally or self-host the open-source project, while teams wanting managed storage, issue detection, and collaboration can use Arize AX. This makes Phoenix attractive to builders who want an instrumentation path that does not begin with a mandatory SaaS contract.
Arize's September 2026 pricing page lists AX Free at $0 with 25,000 trace spans per month, 1 GB monthly ingestion, 15-day retention, unlimited users, and unlimited evaluations. AX Pro costs $50 per month and raises the allowance to 50,000 spans, 10 GB ingestion, and 30-day retention while keeping unlimited users and evaluations. Enterprise pricing is custom, with unlimited issue signals, custom spans, ingestion, and retention, plus SaaS or self-hosted deployment. Open-source Phoenix can be self-hosted without a license fee, although infrastructure, backups, upgrades, and operational labor remain real costs.
Phoenix works well when a team needs to answer specific questions: Which tool call failed? Did a prompt change improve groundedness? Which retrieved documents caused a bad response? Did latency or token use regress? A disciplined workflow captures traces, creates a dataset from representative production examples, defines evaluators, and compares experiments before deployment. For high-volume workloads, estimate spans rather than requests—a single agent run can emit many spans and exhaust 25,000 or 50,000 faster than expected. Teams should also decide whether 15 or 30 days of managed retention is sufficient for incident review.
Alternatives include AgentOps for agent-centered session debugging, LangSmith for teams deep in LangChain, Langfuse for open-source LLM engineering, and Braintrust for evaluation-led iteration. Phoenix's strengths are standards-based traces, open-source access, and a credible path to managed AX. Its gaps are that self-hosting requires engineering ownership, costs become custom beyond published allowances, and native MCP support was not stated on fetched pages. Pilot one production chain with 100 representative examples and explicit quality, latency, and cost thresholds before expanding instrumentation.
Evaluation scores also need calibration. Start with human-labeled examples, measure evaluator agreement with reviewers, and inspect false positives and false negatives before using a score as a deployment gate. Store prompt, model, retrieval configuration, and code version with each experiment so improvements are reproducible. OpenTelemetry compatibility helps, but teams still need a consistent span taxonomy and redaction policy. Without those conventions, traces from different services become difficult to compare and may expose customer data.
Leading open-source LLM observability platform offering comprehensive tracing, evaluation, and experimentation without vendor lock-in. Ideal for teams with DevOps capacity who need deep analytical insights into LLM application behavior, RAG pipeline quality, and multi-agent workflow debugging. Phoenix stands out for its OpenTelemetry foundation, which ensures trace portability and avoids ecosystem lock-in, and its robust evaluation framework that supports both automated LLM-as-a-judge scoring and human annotation workflows. The self-hosted model with zero licensing costs makes it particularly attractive for regulated industries and cost-conscious teams, though the operational overhead of managing infrastructure and the steeper learning curve compared to polished SaaS alternatives like LangSmith should be weighed against these benefits. With over 18,000 GitHub stars and strong backing from Arize AI, the project demonstrates sustained momentum and community adoption.
Free
$0
$50/month
Contact sales
Ready to get started with Arize Phoenix?
View Pricing Options →Arize Phoenix works with these platforms and services:
We believe in transparent reviews. Here's what Arize Phoenix doesn't handle well:
Through late 2025 and into 2026, Phoenix has expanded agent-focused tracing with deeper support for LangGraph, CrewAI, and AutoGen, including visualizations for multi-agent coordination and tool-call sequence inspection. The evaluation framework has been enhanced with new built-in evaluators for code generation quality, multi-turn conversation coherence, and structured output validation. Session and thread-based tracing now provides better visibility into conversational AI applications, grouping related interactions and tracking context evolution across turns. The prompt playground has been upgraded with multi-model comparison capabilities, allowing teams to test prompts against several providers simultaneously and feed results directly into experiments. Guardrails integration enables teams to define and monitor safety boundaries alongside performance metrics. The annotation workflow has been streamlined with bulk labeling tools, inter-annotator agreement metrics, and API-driven integration with external labeling platforms. Infrastructure improvements include faster trace ingestion, improved query performance for large datasets, and better support for high-cardinality span attributes in production environments.
LLM Observability
LangSmith helps developers trace, evaluate, test, and monitor applications built with language models and agent workflows.
LLM Observability
Open-source LLM engineering platform that captures traces, runs evaluations, and helps teams analyze cost, latency, and quality across production AI apps.
AI Evaluation
Braintrust review 2026: unified evals, prompt workbench, observability, and AI proxy used by OpenAI, Stripe, Notion — features, pricing, pros, cons.
AI Monitoring
Open-source LLM observability and gateway platform offering one-line integration, request logging, prompt management, evals, and a multi-provider proxy.
MLOps
End-to-end MLOps and AI developer platform — Models (experiment tracking, sweeps, model registry) plus Weave (LLM/agent observability and evals) — used by frontier labs and enterprise ML teams.
Pricing and features change often. If anything on this Arize Phoenix page is wrong or missing, tell us and include a link to the source. We review every correction.