
Excels at experiment tracking and model registry workflows, but is less purpose-built for real-time LLM production observability.

Arize AI records 249,028 monthly visits with +3.2% month-over-month growth and Paid, Contact for Pricing pricing. Roark records 9,226 monthly visits with +23.1% growth and Paid pricing.

Test, monitor and improve the voice AI agents you already run
| +1.73K (+23.1%) |
| Primary Market & Share | 🇺🇸 United States (51.6%) | 🇺🇸 United States (75.7%) |
|---|
| Direct Visits / mo | 107.84K | 4.83K |
|---|
| User Rating | 4.6 | 4.4 |
|---|
| Supported Platforms | Website | Website |
|---|

Agent observability, evaluation, and continuous improvement for production AI
Capture spans, traces, and sessions across agents and LLM calls using OpenInference conventions built on OpenTelemetry, so instrumentation stays portable.
Run evaluations at span, trace, and session level to grade retrieval quality, tool use, and final answers across large production datasets.
An in-product agent that inspects traces, diagnoses likely root causes, and proposes fixes that engineers can review before shipping.
Self-host the leading open-source tracing and eval tool when you need to keep trace data inside your own infrastructure.

Test, monitor and improve the voice AI agents you already run
Run your agent against hundreds of realistic personas, accents, background noise and edge cases before a customer ever dials in.
Red-team callers attempt prompt injection, jailbreaks and social engineering to check your agent holds policy under pressure.
Every live call is scored as it lands on more than 500 metrics, with issues filed, alerts fired and traces attached.
Purpose-built audio models measure pronunciation, emotion, vocal stress, pace, interruptions and dead air, not just transcript quality.

Excels at experiment tracking and model registry workflows, but is less purpose-built for real-time LLM production observability.


Free / Open-sourceBuild, simulate, evaluate and monitor self-improving AI agents
#2
Free / Open-sourceEvaluate, monitor, and govern your AI applications in production

Focuses heavily on prompt management and versioning control, whereas Arize prioritizes end-to-end tracing and ML observability.

Free / Open-sourceStandardize LLM evaluation, observability, and red teaming across your whole org

Automated QA, simulation and observability for voice and chat AI agents

Free / Open-sourceAgent observability, evals and governance for AI platform teams

Free / Open-sourceAI governance, testing, and observability for regulated enterprises

Free / Open-sourceOpen-source observability that finds your AI agent's failures — and fixes them

Free / Open-sourceOpen-source workspace for building, evaluating and shipping AI agents

Provides enterprise governance and model explainability features, but has a steeper setup curve for pure LLM tracing.

Free / Open-sourceFind and fix problems in your AI agents before users do
Excels at experiment tracking and model registry workflows, but is less purpose-built for real-time LLM production observability.
Free / Open-sourceBuild, simulate, evaluate and monitor self-improving AI agents
Free / Open-sourceEvaluate, monitor, and govern your AI applications in production
Focuses heavily on prompt management and versioning control, whereas Arize prioritizes end-to-end tracing and ML observability.
Free / Open-sourceStandardize LLM evaluation, observability, and red teaming across your whole org
Free / Open-sourceAgent observability, evals and governance for AI platform teams
Free / Open-sourceAI governance, testing, and observability for regulated enterprises
Free / Open-sourceOpen-source observability that finds your AI agent's failures — and fixes them
Free / Open-sourceOpen-source workspace for building, evaluating and shipping AI agents
Provides enterprise governance and model explainability features, but has a steeper setup curve for pure LLM tracing.
Free / Open-sourceFind and fix problems in your AI agents before users do