Copilot Usage Monitor
Understand how much of your coding is AI-assisted (GitHub Copilot) vs manual — session trends, tokens when available, and optional reports you can share with your manager.
Get visibility into Copilot adoption from editor activity and optional Copilot log correlation. Everything is stored locally on your machine until you export it.
Install: Visual Studio Marketplace · VS Code 1.85+ · GitHub Copilot + Copilot Chat
Important: This extension does not call GitHub’s official Copilot Usage Metrics API. Numbers are heuristic estimates, not billing or org telemetry. For org-wide dashboards, see Microsoft Copilot Metrics Dashboard.
Real-time & historical metrics
Track how you code with Copilot in the background:
- AI vs manual ratio — by characters or lines (Tab accepts, Chat/Composer-style applies, manual typing)
- Current session — live percentage in the status bar
- Today / 7 days / 30 days — rollups with event counts
- By language, model, and project — when enough data and logs are available
- Token estimates — from Copilot logs when present, with optional character-based fallback
- Copilot vs other AI-shaped edits — when log correlation marks
provider: copilot
Open the Usage Report from the activity bar (Copilot Usage) for tables, coaching tips (private to you), and refresh anytime.
Manager-friendly sharing (optional)
Nothing is sent to your manager or GitHub automatically.
- Use Copilot normally for a week or two (Tab, Chat, Composer).
- Review the Usage Report sidebar.
- Export when you choose:
Copilot Monitoring: Export Manager Report — Markdown or HTML (print to PDF)Copilot Monitoring: Copy Teams Summary — one line for Teams or email
- Share the file only if you want to.
Your name on exports: set copilotMonitoring.displayName, or use signed-in GitHub username, or your OS account name.
|
Copilot Usage Monitor |
Org Copilot Metrics Dashboard |
| Who |
Individual developer |
Platform / org admin |
| Data |
Local heuristics + optional Copilot logs |
Official GitHub Copilot Usage Metrics API |
| Sharing |
You export voluntarily |
Admin grants dashboard access |
Getting started
- Install from the Marketplace.
- Enable GitHub Copilot and GitHub Copilot Chat and sign in.
- Edit code as usual — tracking runs when
copilotMonitoring.enabled is on (default true).
- Click the status bar AI % or open Copilot Usage → Usage Report.
- Optionally run Start Calibration Wizard to verify Tab vs Chat classification in a scratch workspace.
Zero configuration (defaults)
- Monitoring on at startup
- Copilot log tailing on (for provider, model, and token hints)
- 90-day retention, up to 50,000 events
- No API keys required for core features
Privacy & data location
- 100% local — events live under VS Code global storage until you export or clear data.
- No cloud upload by this extension — no prompts or full chat transcripts stored.
- You control the data — open the storage folder, export, or Clear All Data anytime.
Typical storage folder:
- Windows:
%APPDATA%\Code\User\globalStorage\KumVivek.copilot-usage-monitor\
- macOS:
~/Library/Application Support/Code/User/globalStorage/KumVivek.copilot-usage-monitor/
- Linux:
~/.config/Code/User/globalStorage/KumVivek.copilot-usage-monitor/
Files: events.ndjson (classified edits), chat-signals.ndjson (log-derived chat metadata summaries).
Commands
Command Palette (Ctrl+Shift+P / Cmd+Shift+P):
| Command |
What it does |
Copilot Monitoring: Show Usage Report |
Open the sidebar report |
Copilot Monitoring: Show Dashboard |
Full dashboard webview |
Copilot Monitoring: Export Manager Report |
Curated report for your manager |
Copilot Monitoring: Copy Teams Summary |
Short summary to clipboard |
Copilot Monitoring: Export Stats |
CSV / Markdown rollups |
Copilot Monitoring: Open Event Storage Folder |
Reveal local data folder |
Copilot Monitoring: Start Calibration Wizard |
Checklist + log samples |
Copilot Monitoring: Clear All Data |
Delete local events and chat signals |
Copilot Monitoring: Set GitHub Metrics Token |
Optional org metrics on export |
Configuration (optional)
Open Settings and search copilotMonitoring:
| Setting |
Default |
Purpose |
enabled |
true |
Master on/off |
enableCopilotLogTailing |
true |
Attribute Copilot provider / model / tokens from logs |
displayName |
"" |
Your name on manager reports |
defaultReportPeriodDays |
30 |
Period for Teams summary and coaching tips |
detectionProfile |
standard |
Use relaxed for lower Chat thresholds when logs are missing |
rollupTimezone |
utc |
local for day/week buckets |
retentionDays / maxEvents |
90 / 50000 |
Local data retention |
weeklyReminderEnabled |
false |
Optional weekly export reminder |
showGitHubLoginOnReport |
true |
Show @github on exports when signed in |
enableGitHubMetricsEnrichment |
false |
Optional org API block on manager export (needs permission) |
Threshold tuning: minInsertionChars, cursorJumpWindowMs, enableAgentApplyDetection, token estimation options — see Settings descriptions.
Perfect for
- Developers — see your own AI vs manual mix and improve how you use Tab vs Chat
- 1:1s and reviews — voluntary export of adoption, not surveillance
- Teams piloting Copilot — lightweight self-report alongside official org metrics
Limitations (read before you trust the numbers)
- Estimates only — paste, small Chat applies, or missing logs can skew AI %.
- VS Code + Copilot — built for Microsoft’s VS Code log layout (
Code/logs).
- No full chat text — VS Code does not expose Copilot Chat transcripts to extensions; exports use rollups and summaries.
- Model may be unknown — depends on Copilot Chat log formats; formats can change between Copilot updates.
- Character-based ratio — default is chars added; switch to lines in the Usage Report toolbar.
More detail: KNOWN_GAPS.md in the extension package.
License
MIT — see LICENSE in the repository.
