功能对比
| CSV field | Accepted names | Used for |
|---|---|---|
| User | user, username, login, actor, email | Team member grouping and overage review. |
| Model | model, model_name, sku | Model-level AI credit or premium request analysis. |
| Date | date, day, created_at, timestamp | Daily burn-down and migration windows. |
| Premium requests | premium_requests, requests, request_count | Legacy request-style tracking. |
| AI credits | ai_credits, credits, credit_usage | Usage-based budget tracking. |
| Cost | cost, usd, amount | Fallback when exports include dollars but not credits. |
Local Analyzer
Use the tool below to paste or upload CSV. Parsing happens in the browser session; the page does not send the file to a server. Last updated July 1, 2026; GitHub billing exports and field names can change.
What The Analyzer Reports
The analyzer groups rows by user, model, and date, then estimates total premium requests, AI credits, daily burn, and overage risk against a team allowance.
- User summary: who is consuming the most premium requests or credits.
- Model summary: which models drive the budget.
- Date summary: burn spikes and billing-cycle risk.
- Budget worksheet: allowance, used credits, remaining credits, projected overage risk.
Privacy Statement
The CSV parser is client-side React state. Avoid pasting secrets, tokens, or private customer data into any CSV audit workflow, and keep official billing exports in approved storage.
- No upload endpoint is used by this page.
- File input is read with the browser FileReader API.
- Results disappear when the page is refreshed unless copied elsewhere.
- Verify final billing numbers inside GitHub billing settings.
Projected monthly usage may exceed allowance.
Detected fields: user user, model model, date date, requests premium_requests, credits ai_credits.
Projected 30-day credits from parsed daily average: 6,555. This is a planning estimate, not an official GitHub bill.
| User | Requests | AI credits | Rows |
|---|---|---|---|
| alice | 40 | 242 | 2 |
| bob | 9 | 140 | 1 |
| carol | 45 | 55 | 1 |
| Model | Requests | AI credits | Rows |
|---|---|---|---|
| gpt-5-codex | 40 | 242 | 2 |
| claude-sonnet-4.5 | 9 | 140 | 1 |
| gpt-5-mini | 45 | 55 | 1 |
常见问题
Does this Copilot usage CSV analyzer upload my file?
No. The analyzer reads uploaded or pasted CSV text in the browser and stores results in local component state only.
What fields does the analyzer need?
It works best with user, model, date, premium request, AI credit, or cost fields. It also accepts common variants such as username, actor, timestamp, credits, or usd.
Can this replace GitHub billing?
No. Use it for audit and planning, then verify official usage, budgets, and invoices in GitHub billing.
How do I estimate overage risk?
Enter a team monthly credit allowance. The analyzer compares parsed AI credits or estimated credits against that allowance and projects risk from daily burn.