The Agent Execution Feed helps you monitor how agents are performing, what models they’re using, and how efficiently they’re running. It gives you the insights you need to improve performance, manage costs, and debug issues.

​
What You’ll See in the Feed

Each row in the feed includes:

  • Model Provider – Who’s powering the model (e.g., OpenAI, Anthropic).
  • Model Name – The specific model used.
  • Project – The project the agent belongs to.
  • Source – What triggered the agent (e.g., API, UI).
  • Agent Name – The specific agent that ran.
  • Response Time – How fast the agent responded.
  • Total Tokens – Tokens used during execution (helps with cost tracking).

A higher response time could indicate potential issues in the agent’s performance or inefficiencies in processing that need to be addressed.

​
Drill into Execution Details

Click on a row to view more:

  • User Input – What the agent received.
  • Returned Chunks – Any retrieved context.
  • Agent Output – The final response generated.

This information can be used as an important audit tool to review any agent execution for performance and accuracy.

​
Filter for Deeper Insights

Narrow your view by:

  • Date
  • Project
  • Source

These filtering options allow for targeted analysis, helping you drill down into specific aspects of agent performance and optimize where necessary.

​
Refresh & Export

  • Refresh to get the latest data in real-time.
  • Export to CSV to analyze offline, share with your team, or connect with other tools.