Zest

AI CODING ANALYTICS

GitHub Copilot usage
metrics, per engineer.

For GitHub Copilot metrics, your Copilot settings show your usage this cycle in AI credits. Org and enterprise owners get a usage metrics dashboard under Insights and an API with a daily report per user; Zest adds Copilot Chat next to Claude Code, Cursor and Codex, with no demo call.

Start free
npx @winding-labs/zest

Free for 3 seats · No credit card · One command signs in every agent on your machine

Sample data

Stack in the Zest app: how well your setup is wired, and what raises the score

USED BY ENGINEERING TEAMS AT

Flyps Charm Wildfires Geoforge MisaLabs

How to check your own Copilot usage

Copilot counts usage in AI credits. Where you see yours depends on the plan:

  • On a Copilot Business or Enterprise seat: click your profile picture, then Copilot settings. Under Usage, Usage this cycle shows the credits you've used.
  • On a personal plan: github.com/settings/billing, then AI usage, with the credits each model used and their cost.
  • In the Copilot CLI, /usage shows this session's tokens per model and, on token-based billing, each model's AI credits.

Are GitHub's own Copilot reports enough?

For a team that only uses Copilot, yes. GitHub's dashboard and per-user reports count completions, chat modes and AI credits, and Zest doesn't capture inline completions at all. They stop being enough when the same engineers also run Claude Code, Cursor or Codex on their own plans, because GitHub only reports third-party agents that run as agent apps on GitHub.

Zest is worth installing when your team mixes Copilot Chat in VS Code with other coding agents and keeps its code on GitHub. It won't help a Copilot-only team that wants acceptance rates, anyone on Copilot in JetBrains or Visual Studio, which the extension doesn't read, or a team whose code lives outside GitHub.

We read GitHub's column in its docs on November 10, 2026; Zest's column comes from its setup guide.
QuestionGitHub's Copilot usage metricsZest
Per engineerYes, in the users-1-day and users-28-day reportsYes, session by session in Copilot Chat
Inline completionsYes, with acceptanceNot captured
ChatPrompt counts and totals by feature and modelThe conversation, the model, the tokens Copilot reports and each tool called
Where it readsIDEs and Copilot CLI in the reports; the dashboard charts leave CLI outCopilot Chat in VS Code
Pull requestsDaily totals in the enterprise, organization and repository reports, and by adoption phase on the impact dashboard; none in the per-user reportsEach engineer's sessions next to the pull requests they merged
Other coding agentsRecognized agent apps on GitHub, in public previewClaude Code, Codex, Cursor, Grok Build, Hermes, OpenCode and Freebuff, each through its own plugin or extension
Who sets it upAn enterprise or organization owner turns on the metrics policyEach engineer installs the extension; an admin installs the GitHub App
FreshnessWithin two full UTC days after the day closesThe extension syncs every minute

Where is the GitHub Copilot usage metrics dashboard?

Open your enterprise or organization, then Insights, then Copilot usage; the Copilot usage metrics policy has to be on first. Enterprise and organization owners see it, and so do billing managers and custom roles with the View Copilot Metrics permission. Its three dashboards use a 28-day window, the current calendar month or seven days, depending on the metric:

  • Copilot usage: daily and weekly active users, code completions and their acceptance rate, requests per chat mode (ask, edit, plan, agent), and model and language use. Copilot CLI isn't in these charts.
  • Code generation: lines changed with AI, and the share that agents added or deleted.
  • Copilot impact: users grouped into adoption phases, compared on pull requests merged per user per month and time to merge.

Is there a GitHub Copilot metrics API?

Yes. The REST API returns the reports as files. Each endpoint under /orgs/{org}/copilot/metrics/reports/ (or /enterprises/{enterprise}/…) answers with signed links to NDJSON reports: one day or the latest 28 days for the organization or per user, and one day only per repository or per user and team. An organization token needs read:org, or the Organization Copilot metrics read permission. A day's data is available within two full UTC days after it closes; reports start on October 10, 2025, and one-day reports reach back up to a year.

Each per-user line has user_login and day, prompts sent, code generated and accepted, lines added and deleted, flags such as used_agent, used_chat and used_cli, ai_credits_used, and totals by IDE, feature, language and model. GitHub says ai_credits_used is for consumption analysis, not invoicing totals. Per-user lines have no pull requests; those are in the aggregate and repository reports. The old /orgs/{org}/copilot/metrics endpoint closed on April 2, 2026.

curl -L \
  -H "Accept: application/vnd.github+json" \
  -H "Authorization: Bearer <YOUR-TOKEN>" \
  -H "X-GitHub-Api-Version: 2026-03-10" \
  "https://api.github.com/orgs/ORG/copilot/metrics/reports/users-1-day?day=DAY"

GitHub Copilot metrics don't show Claude Code usage: what does?

GitHub's Copilot metrics only see Claude Code when it runs through Copilot. The reports count job starts for the Claude and Codex agent apps on GitHub (totals_by_3rd_party_agent), and requests and tokens from tools that use Copilot, with claude_code and codex among the client values (totals_by_client). GitHub calls that client value a best-effort grouping.

Claude Code on an Anthropic plan or API key isn't in those reports. For that, use Anthropic's dashboard, Analytics API or OpenTelemetry, all covered on Claude Code usage. To see Copilot and Claude Code per engineer in one place, you need a tool that reads both agents.

What GitHub's Copilot usage metrics count, from GitHub's docs and changelog, read November 8, 2026.
Where the agent runsIn GitHub's Copilot metrics?
Copilot in an editor or the cloud agentYes
Claude or Codex agent apps on GitHubJob starts per agent, no tokens or models
Claude Code or Codex through CopilotRequests and tokens per client
Claude Code on an Anthropic plan or API keyNo
Codex on a ChatGPT planNo
CursorNot a named client

How Zest measures GitHub Copilot usage

GitHub's reports describe Copilot. Claude Code, Cursor or Codex sessions on the same engineer's laptop aren't in them unless they run through Copilot, which is why Claude Code vs GitHub Copilot, Codex vs GitHub Copilot and Cursor vs GitHub Copilot are worth measuring on your own team. Zest answers the cross-tool question: which coding agent, model and skills each engineer reaches for, and which sessions turn into merged pull requests.

  • The Zest extension for VS Code (setup) reads Copilot Chat sessions: the conversation, the models, the tokens Copilot reports, and the skills and tools each one called.
  • Inline completions are not captured. GitHub's own metrics stay the source for completion acceptance.
  • Claude Code, Codex, Cursor, Grok Build and Hermes report into the same dashboard, so usage is comparable across tools.
Stack in the Zest app: how well your setup is wired, and what raises the scoreSample data

Which tools tie Copilot sessions to merged pull requests and defects?

To measure Copilot ROI, GitHub's impact dashboard compares adoption phases on merged pull requests. Zest shows it per engineer: the Zest GitHub App puts each engineer's Copilot Chat sessions next to the pull requests they merged, with PR cycle time. GitHub's return-on-investment estimates are in the dashboard only, not in the API or the NDJSON exports.

Zest reads no CI, issue tracker or incident data, so it shows no defects. A 300-engineer org that needs both can compare the engineering intelligence platforms that read those sources. Which delivery numbers to trust is on developer productivity metrics.

TRY IT ON YOUR TEAM

See what your team ships with AI.

Starter is free (3 seats, no credit card). Pro is $49 a month with 5 seats and Max is $149 a month with 10 seats — priced per workspace, not per engineer, with extra seats at $19 a month. See pricing

Start free
npx @winding-labs/zest

A FEW THINGS TO KNOW

Questions

Where do I see GitHub Copilot usage metrics?+

Enterprise and organization owners open Insights, then Copilot usage, once the Copilot usage metrics policy is on. The same data comes as daily NDJSON reports from the /copilot/metrics/reports REST endpoints.

How do I check my own Copilot usage?+

On a Business or Enterprise seat, open Copilot settings and read Usage this cycle, in AI credits. On a personal plan, go to github.com/settings/billing, then AI usage.

Can I see GitHub Copilot usage per engineer?+

Yes. The users-1-day and users-28-day reports have one line per user, with prompts, code accepted, lines added and AI credits. Zest adds each engineer's Copilot Chat sessions next to their other coding agents and merged pull requests.

Is the GitHub Copilot Metrics API deprecated?+

The old /orgs/{org}/copilot/metrics endpoint closed on April 2, 2026. Use the Copilot usage metrics reports under /copilot/metrics/reports instead.

Does Zest capture Copilot inline completions?+

No. Zest reads Copilot Chat sessions in VS Code. For completion acceptance rates, use GitHub's Copilot usage metrics.