Skip to content

Anthropic API cost analyzer

Drop the usage CSV from the Anthropic Console and see your Claude API cost broken down by API key, by model, and by day, computed from the official per-model token rates.

Runs entirely in your browser. Nothing you enter leaves this page.

The file never leaves your browser. No upload, no account.

Drop your Anthropic usage CSV here

or click to choose a file. Parsed locally, never uploaded.

How to get your usage export

  1. Open the Anthropic Console and go to Usage and Cost.
  2. Export the per-key token usage as CSV.
  3. Drop it above. It is parsed in your browser; nothing is uploaded.

Rates verified on 2026-07-10 and reproduced against a real export. Cost is the standard list-rate value of your token usage; your invoice may differ on custom rates.

Frequently asked questions

Is anything uploaded?

No. The CSV is read and parsed with JavaScript in your browser tab. There is no upload endpoint and no account. You can verify in your browser dev tools or disconnect from the network after the page loads.

How is cost calculated?

Per model, from the token columns in the export: uncached input at the model rate, cache reads at 0.1x input, 5-minute cache writes at 1.25x input, 1-hour cache writes at 2x input, and output at the output rate. Rows marked as batch usage are halved. The result is the standard list-rate cost of that usage.

Why might this differ from my Anthropic invoice?

It uses published list rates verified on the date shown below the tool. Differences come from any custom or committed-use rates on your account, rate changes since that date, or services not in the token export (for example web search, billed separately). Current Claude models include the 1M-token context at standard pricing, so the context-window column does not change the rate.

Where do I get this CSV?

In the Anthropic Console, open Usage and Cost and export the per-key token usage CSV. Its columns are usage_date_utc, model_version, api_key, workspace, the four input-token columns, and usage_output_tokens.

What does per-API-key attribution give me?

It shows which key, and therefore which app, environment, or teammate, drove the spend. That is the breakdown the Console summarizes but does not chart over time. Goei keeps it continuous across keys and providers.

Can I share a screenshot without leaking API key or workspace names?

Yes. Turn on "Anonymize names for screenshots" above the results and every API key and workspace label is replaced with a neutral alias (Key 1, Workspace 1, and so on), while the dollar, token, and share values stay exactly as shown. Nothing is sent anywhere either way; this only changes what the table displays.

Sources

Prices last verified 2026-07-10, from each provider's official page.

Independent tool, not affiliated with Anthropic. Estimates at list rates; your bill is set by Anthropic.

Want your whole stack in one number, subscriptions included? Total it with the How much are you spending on AI tools?.