Skip to main content
Arize AX is an AI engineering platform for improving AI agents and applications by observing real behavior, evaluating quality, and proving each change works before you ship. Arize AX gives you the tools to evaluate agent trajectories and decisions with real-time monitoring and automated alerting to surface failures the moment they happen. Sign up for free to get started, then follow the Arize AX workflow below to capture traces, evaluate the quality of outputs, and improve across prompts, models, and agent logic.
Signal in Arize AX, with surfaced issues and issues dashboard

New to Arize AX?

Start here. Instrument your agent, view your first trace, and create your first eval.

​
The Arize AX workflow

1

Instrument

Capture your first trace. Instrumentation adds tracking to your agents to capture traces: step-by-step records of agent activity (inputs, outputs, invoked tools, and costs). After instrumentation, your application streams traces into Arize AX to lay the foundation for measurable improvement.
Using Claude, Cursor, or Codex? Paste Follow the instructions at https://arize.com/docs/PROMPT.md and ask questions as needed into your coding agent to install and instrument with Arize AX in minutes.

Coding agent setup

Use AI coding agents to easily set up Arize AX via skills, MCP servers, or a single URL.

Auto-instrumentation

Instrument your LLM application with pre-built integrations for 30+ providers and frameworks

Tracing concepts

Understand traces, spans, and how they capture each run.
2

Observe

Identify where your agent is failing. Once traces are flowing, use Signal to surface issues, or ask Alyx to surface failing runs in plain language. Look across multiple runs to spot patterns, and create dashboards and alerts to maintain visibility.

Explore your traces

Search, filter, and dig into traces to find what went wrong.

Get started with Signal

Automatically surface patterns and recurring issues with Signal.

Find problematic traces

Use queries, filters, and AI Search to surface bad experiences.
3

Evaluate

Create repeatable checks for quality. Review traces and categorize outputs with annotations. Good examples become a curated collection of test cases, while failures reveal mistakes that can be identified with evaluators: repeatable tests that score outputs against set criteria.

Create evaluators

Set up automated evals so every response is scored without manual review.

Annotate your outputs

Annotate data to identify failure patterns.

Align evaluators

Calibrate automated scores against human judgment.
4

Improve

Test changes and measure their impact. Change the prompt, swap the model, or adjust how the app retrieves information, then run it against your test cases to verify improvement. Each run is an experiment, and Arize AX tracks every version so you can compare them and see which performs best.

Improve your agent

Turn production failures into fixes and prove them across your dataset.

Build a dataset

Curate test cases that capture how your app should behave.

Set up an experiment

Run a change against your dataset and measure the impact.
Automate every step with AI using Alyx, the AI engineering agent built into Arize AX, the Arize AX CLI inside your coding agent, and the Agent Improvement Loop to instrument, observe, evaluate, and improve.

​
Alyx: Your AI engineering agent for every step

Alyx is the AI engineering agent built into Arize AX. Ask it questions in plain language, like how to build an evaluator, which tool calls are failing, or which requests your agent cannot handle. Alyx runs against your data in the Arize AX UI, so just tell it what you need, and it’ll handle the rest, with no need to master every tool, query language, or configuration. Learn more in the Alyx documentation.

​
Next Steps

Using Claude, Cursor or Codex?

Paste Follow the instructions at https://arize.com/docs/PROMPT.md into your coding agent to install and instrument with Arize AX, or use agent skills for the full workflow.

Get Started with Arize AX

Get started with the instrument, observe, evaluate, and improve workflow for your AI product.

Review pricing and tiers

Compare Arize AX plans and find the tier that fits your needs.

Need OSS?

Check out our OSS product, Phoenix.

​
Resources

Explore our guides

Check out a comprehensive list of example notebooks for agents, RAG, voice, tracing, evals, and more.

Join our Slack community

Join the Arize Slack community to ask questions, share findings, provide feedback, and connect with other developers.
Arize AX is trusted by teams building agents at: Logos of teams and organizations that use Arize AX