usage-insights — GitHub Copilot plugin: what it does, how to install
In-canvas dashboard tracking local GitHub Copilot token metrics, AI credits, and sub-agent call durations.
copilot plugin install usage-insights@awesome-copilotWhat usage-insights does
The extension renders inside GitHub Copilot's canvas to report local AI credits alongside input, output, reasoning, and cache tokens. It aggregates usage across recent sessions, highlights high-cost interactions, and draws a 365-day activity map for quick reviews.
Under the hood, it queries local session stores using Node's built-in node:sqlite without extra package dependencies. Because it binds its loopback service strictly to 127.0.0.1 and opens SQLite files read-only, telemetry stays entirely on your machine.
What it helps with
- inspect AI credit consumption across individual Copilot sessions
- break down token counts between root and sub-agents
- review model call durations across chosen time ranges like 7d or 30d
- view a 365-day token activity heatmap without sending data to external tools
Example prompts for the agent
- Show my Copilot token usage breakdown for the past 7 days
- Open session metrics in canvas to inspect high-cost sub-agent calls
How to install usage-insights
Command from the source
copilot plugin install usage-insights@awesome-copilotStep by step, with commands for each platform: Claude Code plugins: marketplaces, installing, and building your own
Questions about usage-insights
How do I install usage-insights in GitHub Copilot?
Run copilot plugin install usage-insights@awesome-copilot in your terminal.
What does the usage-insights plugin do?
It surfaces session metrics, agent breakdowns, and a 365-day token heatmap directly within Copilot's canvas view.
Does usage-insights transmit session logs off the machine?
It binds exclusively to 127.0.0.1, accesses SQLite databases read-only, and sends zero metrics to remote endpoints.
