tools
Arize AX
Arize AX traces, evaluates, and improves AI agents and applications using observability, issue detection, experiments, and managed agents.
Arize AX is an AI engineering platform for capturing application traces, finding recurring failures, evaluating quality, and testing changes before deployment. It supports AI engineers and product teams working on agents and other AI applications.
The platform includes tracing, Signal issue detection, evaluators, datasets, experiments, prompt workflows, and Alyx. The free plan has usage limits; managed-agent workflows, repository access, self-hosting, and enterprise controls require higher-tier plans.
Features
- Capture traces for inputs, outputs, tools, sessions, and costs
- Use Signal to group recurring failures into ranked issues
- Search and filter traces with queries and AI Search
- Evaluate traces and spans with LLM, agent, and code evaluators
- Run experiments comparing prompts, models, or agent changes
- Build datasets and regression tests from production behavior
- Use Alyx for natural-language AI engineering assistance
- Access APIs, SDKs, CLI commands, and OpenTelemetry integrations
Use cases
- Trace production AI agents and inspect their tool calls
- Find recurring failures and investigate their likely causes
- Score agent responses against quality and safety criteria
- Compare prompt, model, retrieval, or agent changes
- Create regression datasets from failing production traces
- Monitor token usage, latency, and model costs
Pros
Cons
Latest updates
- Get a webhook when an eval task fails
Subscribe a webhook to ONLINE_TASK_RUN_FAILED to hear when an eval task stops writing evals.
- Fixes and improvements
Experiments are now marked stable in the SDKs; fixes cover proxy authentication, dashboards, custom metrics, annotations, and Alyx.
- Run managed agents on the OpenCode harness
OpenCode joins Claude Code, Codex, and Cursor as a harness for managed agents, and it is now available on every account.
- Find and fix eval gaps with the Eval Health agent
A managed agent reviews the evals on a tracing project and proposes changes for you to accept.
- Score whole conversations with trace and session remote evaluators
A remote evaluator receives each mapped variable as an array with one value per span, so your endpoint can score an entire conversation in one call.
Capabilities
Security
- SOC 2 Type II — “Arize AI Announces SOC 2 Type II Certification” source
- SOC 2 — “Arize AI Announces SOC 2 Type II Certification” source
- ISO 27001 — “ISO/IEC 27001 Certified” source
- GDPR — “GDPR Compliant” source
- HIPAA — “HIPAA Compliant” source
- Data kept in the EU — “Arize AI: Support for EU Data Residency” source
Pricing
- Starting price
- $50/mo
- Prices checked
- 2026-10-25
AX Free
Free
- 10 issues per month
- 25k spans per month
- 1 GB per month
- 15 days retention
- SaaS deployment
- Unlimited users
AX Pro
- $50 per month
- 25 issues per month
- 50k spans per month
- 10 GB per month
- 30 days retention
- SaaS deployment
- Unlimited users
AX Enterprise
Price on request
- Unlimited signal issues
- Custom span volume
- Custom storage
- Custom retention
- SaaS or Self-Hosted deployment
- Unlimited users