Skip to content
BrainRoad BrainRoad
Documentation Menu

Activity Overview

See what your agent is doing in real time with the Activity feed.

On this page

What the Activity Tab Shows

The Activity tab on your agent’s detail page is a live feed of everything your agent is doing. It streams log output directly from your agent’s container in real time, so you can see exactly what is happening at any given moment.

This is not a summary or a delayed report. It is a direct window into your agent’s runtime behavior, updated as events occur.

Types of Events

The Activity feed captures several categories of events:

  • Messages received — when someone sends a message on WhatsApp, Signal, iMessage, or email, you see it appear in the feed as the agent processes it.
  • Actions taken — replies sent, web pages browsed, files created, code executed. Every action the agent performs is logged.
  • Channel activity — connection status changes for your messaging channels, including reconnections and authentication events.
  • Errors — if something goes wrong (a failed API call, a channel disconnect, a rate limit), it shows up here immediately.
  • Approvals pending — actions requiring human approval appear with a “Pending” status, waiting for your decision before the agent proceeds.

Filtering by Category

The Activity feed offers filtering by event category to help you focus on specific types of activity:

  • All — show all events (default)
  • Communication — messages received and sent across channels
  • Tool Use — agent tool execution and integrations
  • Coding — code generation and execution
  • Lifecycle — agent startup, shutdown, and state changes
  • Social — social media activities
  • Config — configuration changes
  • Errors — errors and failures only

Click any category button to filter the feed to that type of activity. This helps you spot patterns or troubleshoot specific areas quickly.

Filtering by Date Range

You can select a custom date range to view activity from a specific period. Use the date range picker to narrow down to hours, days, or weeks of activity.

AI-Generated Briefings

The “Summary” tab shows an AI-generated briefing of your agent’s activity for the selected time period. The briefing highlights:

  • Key actions completed
  • Significant errors or issues
  • Summary of messages processed
  • Overall agent behavior and patterns

Click Generate to create a briefing for the current view, or refresh an existing one with the Refresh button.

Approvals and Decision Points

Some agent actions require human approval before proceeding. These appear in the Activity feed with a Pending status.

Examples of actions that may require approval:

  • Large or sensitive operations (financial transactions, bulk actions)
  • Actions that don’t match expected patterns
  • High-consequence decisions based on your custom agent rules

When you see a pending approval, click it to review the details and choose to Approve or Reject the action. Your decision is logged and the agent continues accordingly.

Real-Time Updates

The Activity feed uses a WebSocket connection to stream logs as they happen. When you open the tab, you see the most recent 200 lines of history, then new entries appear automatically as the agent produces them.

You do not need to refresh the page. If your connection drops temporarily, the feed reconnects automatically and picks up where it left off.

The Activity feed is read-only. You are observing your agent’s output, not interacting with it. If you need to run commands inside your agent’s environment, use the Console instead.