← All products

Observability and evaluations for AI applications. Inspect traces, investigate failures, and turn real examples into quality tests.

Inside the product

Follow every execution and inspect the details of each step.
See the duration and sequence of steps in the timeline.
Inspect structured inputs across your workflows.
Understand the criteria and results behind quality evaluations.

Understand what your AI is doing

Datool brings traces, evaluations, and quality metrics together. Follow a request from its input to its response and identify where it needs attention.

Inspect each execution

Review inputs, outputs, spans, latency, and scores in the context of the complete workflow. See where an unexpected result came from.

Turn examples into evaluations

Build datasets from real cases, run repeatable evaluations, and compare versions. Combine code-based and language-model scorers with human reviews.

Start with your project

Open Datool, select a workspace and project, and follow the setup instructions to connect your application. Use the first traces to guide your next improvement.