Logs
See every tool call your agent makes, with its input, output and errors
Logs records every tool call your agent makes through ClawLink. Use it to check what ran, find out why a call failed, and see how busy each app is. (Older versions of the dashboard called this page Usage.)
The charts
Tool calls
Calls over the past 7, 30 or 90 days.
Calls by app
Which apps your agent uses most. Failed calls are included here, but only successful calls count toward your plan.
Find a call
Narrow the list with:
- Search: app names, tool names or error text.
- App: one app, or Data for the public data apps.
- Status: Success or Error.
- Time range: the last 24 hours, 7 days, 30 days, or all time.
Open a call
Click a row to see the details:
| Field | What it tells you |
|---|---|
| Status | OK or Failed, with the error message if it failed |
| Device | Who made the call: OpenClaw, Hermes, the CLI, a remote MCP client, or a manual key |
| Input | The arguments your agent sent |
| Output | What the app returned |
| Connection ID | Which connected account ran the call |
| Provider request ID | The app's own id for the request. Give it to the app's support if you need to |
For a public data app, the row also shows the price of the call, or Refunded when the call failed.
When you see errors
Read the error in the call details
The message usually says what went wrong: a missing permission, a bad argument, or a rate limit.
Check the connection
Open Connections and make sure the app shows Connected. If it shows Needs reconnect, reconnect it.
Check the action and its arguments
Make sure your agent called a real action for that app, with real values, not placeholders.
Confirmation needed is not an error. The action changes data, so ClawLink asked your agent to confirm it first. Approve it in your chat and the agent tries again.