Execution History gives you a full record of every time one of your Agents has run — whether it was triggered on a schedule, called from chat, or fired manually. Use it to see what your Agent did, how long it took, and whether it succeeded.
Where to find Execution History
You can access Execution History in three ways:
From the Agents page: Navigate to app.coworker.ai/agents and click the Execution History icon next to any Agent in your My Agents table.
From within an Agent: Open any Agent and select Executions from the top-right actions menu. This filters the view to that specific Agent's runs.
From the Execution pill: Click the black pill on any Agents view to quickly see all recent Agent runs.
The global Execution History page (at /agents/builder/executions) shows runs across all of your Agents.
Reading the Execution History table
Each row in the table represents a single Agent run. The columns show:
Status — the outcome of the run (see statuses below)
Agent — which Agent ran
Started — when the run began
Duration — how long the run took
Summary — a short description of the run or any error message
Drafts — whether the run produced any Draft Actions requiring your review
Click any row to open the Execution Detail drawer, which shows the full chat transcript for that run — including every tool the Agent used, what it searched, what it found, and what output it produced.
Execution statuses
Completed — the Agent finished successfully
Running — the Agent is currently active (the drawer refreshes every few seconds)
Pending — the run is queued and about to start
Failed — something went wrong during the run; the Summary column shows the error
Cancelled — the run was stopped before completing
Credits Exhausted — the Agent hit its monthly credit limit and could not continue. You can increase or remove the credit limit in the Agent's settings.
Filtering and searching executions
Use the filters at the top of the Execution History page to narrow down what you see:
Status filter — show only runs with a specific outcome (e.g., only Failed runs)
Agent filter — limit the view to a specific Agent
Drafts filter — show only runs that produced Draft Actions waiting for review
Mine / Shared toggle — switch between your own Agents and Agents shared with you
The table auto-refreshes every 15 seconds, so you can leave it open while an Agent is running and watch the status update in real time.
Viewing what the Agent did
Clicking a row opens the Execution Detail drawer. Inside, you'll see the full conversation the Agent had with itself as it worked through the task — every search it ran, every tool it called, and every decision it made. For runs that are still in progress, the drawer updates live every few seconds.
If a run produced Draft Actions (write operations held for your approval), you'll see a link to the Action Queue from within the drawer.
Tip: If an Agent run failed unexpectedly, the Summary column and the Execution Detail drawer are the fastest way to understand what went wrong and whether you need to adjust the Agent's instructions or tool access.