One run timeline
Follow the order of prompts, tool calls, responses, and handoffs in a single sequence.
Read the run, not just the result Private preview
A timeline for understanding agent work. Bring tool calls, intermediate outputs, and handoffs into one readable view so you can see how a run reached its result.
01 / The product
Three capabilities planned for the first release of Tracegrove.
Follow the order of prompts, tool calls, responses, and handoffs in a single sequence.
Put two runs side by side to inspect where their inputs, steps, or outputs started to diverge.
Attach observations to a step so another teammate can pick up the investigation with context.
02 / The workflow
Explore the intended workflow, from the first input to a useful result. This is an interactive concept preview.
Follow the order of prompts, tool calls, responses, and handoffs in a single sequence.
Put two runs side by side to inspect where their inputs, steps, or outputs started to diverge.
Attach observations to a step so another teammate can pick up the investigation with context.
03 / Good to know
The planned product focuses on understanding recorded runs rather than operating an agent. Event sources and ingestion formats are still being scoped.
No. All events on this page are illustrative sample data. Tracegrove is a product in development.
A short description of your agent workflow and what makes debugging it difficult is enough. Please do not email private traces or credentials.
Help shape what comes next
Tell us about your workflow and what you need from Tracegrove. We’re opening conversations with potential early users.