Live token count and USD cost for the active prompt file, right in the status bar — across Claude, GPT-4o, Gemini, Mistral, and Cohere.
Most "token counter" extensions stop at counts. Tokenometer adds the dollar value, using the same pricing the tokenometer CLI and GitHub Action use, so what you see in the editor matches what you'd see in CI and on the npm side.
Install
- VS Code: Marketplace listing — or run
ext install faraa2m.tokenometer-vscodefrom the command palette. - Cursor / VSCodium: Open VSX listing.
Build locally (.vsix)
To side-load a development build:
npm install
npm run build --workspace=@tokenometer/vscode
npm run package:vsix --workspace=@tokenometer/vscode
code --install-extension packages/vscode/tokenometer-*.vsix
The same .vsix works in Cursor and VSCodium.
What it shows
The status bar (right side) shows three things, separated by ·:
opus-4-7 · 1,234 tok · $0.0186
- Model — the active model id, shortened (full id in the tooltip).
- Tokens — input token count for the current file. A leading
~means the count is approximate (e.g. Claude / Gemini in offline mode). - Cost — input cost in USD at the model's published per-1k-input rate.
Hovering reveals the full model id, the exact token count, the 8-decimal cost, the tokenizer name, and a hint to click to switch model.
Settings
| Setting | Default | Description |
|---|---|---|
tokenometer.model |
claude-opus-4-7 |
Model id to count and price against. Run Tokenometer: Switch model for the full registry list. |
tokenometer.format |
text |
Format conversion before counting (text, markdown, json, yaml, xml). |
tokenometer.warnOnCostAbove |
0 |
If non-zero, the status bar turns warning-colored when the file's input cost exceeds this USD value. |
Settings can be changed per-workspace (Workspace Settings) or globally (User Settings) — switching via the command saves to the workspace if one is open, otherwise globally.
Commands
- Tokenometer: Switch model — quick-pick over every model in
@tokenometer/core's registry. - Tokenometer: Show details for current file — modal with the full breakdown.
Both commands appear under the Tokenometer: prefix in the command palette.
Supported files
The extension activates only on prompt-shaped content: .md, .markdown, .txt, .json, .jsonc, .yaml, .yml, .xml (matched by language id or extension). Anything else gets a hidden status bar item — no work, no noise.
Files larger than 1 MB are skipped to keep the editor responsive (— shown in place of the count).
How it works
Under the hood the extension reuses @tokenometer/core's tokenize() function — the same path the CLI takes. Counts, pricing, and tokenizer choices are identical to the rest of the suite.
Links
- Root README — methodology, findings, and the full project overview.
tokenometerCLI on npm- Live playground
License
MIT