Skip to main content
Review agent conversations and MCP tool calls to understand what ran, who requested it, and whether it succeeded. Admins manage thread history and Issues from Settings → Ask AI. MCP activity lives under Settings → MCP → Analytics. Each view is scoped to the projects the admin manages. These views record activity. Configure what agents can do through data access controls, SQL mode approvals, and MCP tool curation.

Thread activity

Every agent conversation, in the Lightdash app or Slack, is visible to admins under Threads. It’s the same record the person who asked the question sees: their question, the agent’s answer, and the query behind it. Admins use it to understand how agents are actually being used, and to pull real conversations into Evaluations as test cases.
Threads tab in Ask AI settings showing a list of agent conversations across projectsThreads tab in Ask AI settings showing a list of agent conversations across projects
To see usage across the whole organization, select Insights on the Threads page. It opens the AI Agents Insights dashboard, which summarizes AI agent usage for your organization. Only organization admins can open Insights. Admins who see Threads for the projects they manage can’t open it.

Pull thread data with the API

To load agent conversations into your warehouse or another tool, call the Lightdash API with a personal access token:
  • GET /api/v1/aiAgents/admin/threads lists threads across the projects you manage, with the same scope as Threads. Filter with projectUuids, agentUuids, userUuids, createdFrom, humanScore, dateFrom, dateTo and search. It returns 50 threads per page by default; use page and pageSize to page through.
  • GET /api/v1/projects/{projectUuid}/aiAgents/{agentUuid}/threads/{threadUuid} returns a thread’s messages. Each answer carries its feedback (humanScore, humanFeedback), its token usage (tokenUsage) and its artifacts, whose savedQueryUuid and savedDashboardUuid identify charts and dashboards saved from the thread.
  • GET /api/v1/projects/{projectUuid}/aiAgents/{agentUuid}/threads lists only your own threads with that agent. Add allUsers=true to list everyone’s, which needs permission to manage the agent.

MCP activity

The MCP → Analytics page covers both directions of MCP activity: external assistants calling tools on the Lightdash MCP server, and Lightdash agents calling tools on connected external MCP servers. Admins can see what was called, by whom, and whether it succeeded.
MCP activity feed showing a list of tool calls with tool name, agent, status, and duration columnsMCP activity feed showing a list of tool calls with tool name, agent, status, and duration columns
Activity is retained for 90 days. Which tools an agent is allowed to call is decided separately, per MCP server, in project settings. See choosing which tools are exposed.

Review findings and test fixes

When a conversation reveals a recurring error, use Issues to review findings and proposed fixes. Add representative questions from the thread to an evaluation suite to check whether a change fixes the problem and to catch regressions. Scheduled maintenance has its own record: see Autopilot activity.