Cost Lens for GitHub Copilot

Find out what you're actually spending on GitHub Copilot.

Real AI credits, cost and tokens measured from Copilot's own request logs.

See your spend at a glance

A live badge in the status bar keeps your running credit total in view. Hover for the full breakdown.

A dashboard for every scope

Four tabs Global, Workspace, Session and Optimize each a clean bento grid of where your tokens and money actually go. Open it in the sidebar, or full-screen from the status-bar badge.

Find your priciest prompts

Every prompt ranked by what it really cost in billed credits. Click any one for the full token breakdown, prompt and reply text, tools used, and a per-call table.

Spend less, automatically

The Optimize tab pulls every “spend less” signal into one place: ranked, data-backed tips (long chat history, idle tools, high reasoning left on, and which cheaper model would do the same work for less), a worst-first breakdown of prompting habits detected from your own logs, and a short “how to do better” playbook. Expand any tip for the full breakdown and per-model estimates.

✨ What you get

• 📈 Live status-bar badge your credit total, always visible
• 🗂️ Global / Workspace / Session views with token & cost donuts, openable full-screen
• 💸 Priciest prompts ranked by real billed credits, click for full detail
• ♻️ Cache & tools breakdowns see what rides on every request
• 🧩 Optimize tab ranked tips, detected prompting anti-patterns, and a best-practice playbook
• 💡 Actionable tips including “switch to this model to spend X% less”
• 🔬 Measured, not guessed read from Copilot’s own logs, anchored to billed credits

🚀 Get started

1. Open the Cost Lens view from the activity bar.
2. Click Enable token logging and reload the window.
3. Send a Copilot chat message your real cost appears.

🔒 Private by design

Everything runs locally. The extension only reads logs Copilot writes on your machine and sends nothing anywhere. Prompt and reply text can be redacted with one setting.