Your team deserves more from your agent observability platform - a single, connected place for instrumentation, investigation, and measurement, enhanced with intelligence. In Braintrust, you can start with an open-ended prompt about agent behavior and carry the investigation into code, evals, automation, and monitoring. New tools like Patterns and... Debugger make it easier to find recurring behavior, get to root causes, and act on what you learn. The context you build in one step automatically follows to the next. Work in the platform with an enhanced Loop experience, or move to your coding agent of choice. Read more →show more

Patterns looks through your production traces and identifies recurring failures or opportunities for review. Choose a Slack channel for regular updates, review the findings, and mark each as either useful or noise. Patterns preserves findings over time, and feedback helps future investigations.

Loop lets you interact with anything in Braintrust, including findings from Patterns, in natural language. Ask Loop to investigate across traces, logs, experiments, and datasets in a Braintrust project. Then create Loop automations that repeat open-ended investigations on a schedule. Compare behavior before and after a release, identify accounts adopting a new workflow, or produce a recurring quality review for a product team.

For cost-effective, broad analysis of your traces, use Topics. Topics classifies traces into tasks, issues, sentiment, and custom facets, and turns findings into structured, SQL-queryable context.

Once an issue has been identified, Debugger helps explain why it occurred in a representative trace and provides a diagnosis that can be tested with evals. Move from a symptom to the part of an agent trajectory that needs attention.

All of this works together, with your preferred coding agent, across the same traces, datasets, evaluators, experiments, dashboards, and monitors in your Braintrust environment. Open a Braintrust project, enable Patterns, and review the findings with your team. Use Loop to dig in further, create an evaluator or dataset, and configure monitoring. Get started with the docs →

I like how refreshingly unflashy this video is

🧠🔥

Very cool

This is how we scale intelligence

lfg

🔥

investigation is the real gap, most teams have traces and still cannot answer why the agent did that

The context following a prompt from investigation into code and evals is the interesting part. What exactly gets carried over: evidence, hypotheses, or just a summary?

Trace every tool call and intermediate decision, not just final output; otherwise root-cause analysis misses recovery failures.

gimme the insights!!

yay!!

Connected observability is the right frame — with 20+ agents, the debugging unit is the trajectory, not the span. The platforms that stitch evals, traces and permissions into one surface will own the ops layer of the fleet era.