VS Code Token Counter

A VS Code extension that counts tokens for the active editor using OpenAI tiktoken and shows the result in the status bar.

Install

• VS Code Marketplace: VS Code Token Counter
• Open VSX: vscode-tokencounter

What This Project Does

vscode-tokencount helps you estimate prompt size while writing code or text by showing live token counts in VS Code.

• Shows Tokens: <count> in the status bar for the active editor.
• Recomputes count when you type, switch editors, save files, or change extension settings.
• Supports two counting sources:
  • encoding mode (direct tiktoken encoding, such as cl100k_base)
  • model mode (model mapping, such as gpt-4o-mini)
• Uses adaptive debounce for large files to keep the UI responsive.
• Optionally includes untitled files and can render on the left or right side of the status bar.
• Provides a detailed breakdown (tokens, characters, lines, selected source, resolved encoding) in an output panel.

Only file-backed documents (file:) are counted by default, plus untitled documents when enabled. Non-file schemes (for example git: virtual documents) are intentionally skipped.

Commands

• Token Count: Choose Counter Source (tokenCount.pickCounterSource)
  • Opens a picker to switch between supported encodings and models.
• Token Count: Toggle Detailed Information (tokenCount.toggleDetails)
  • Shows or hides the detailed token info panel.

You can also click the status bar item to open the source picker.

Configuration

All settings live under the tokenCount section:

• tokenCount.countingMode: encoding or model (default: encoding)
• tokenCount.encoding: one of gpt2, r50k_base, p50k_base, p50k_edit, cl100k_base, o200k_base (default: cl100k_base)
• tokenCount.model: one of gpt-4o-mini, gpt-4o, gpt-4.1-mini, gpt-4.1, o3-mini, o4-mini, gpt-5-mini, gpt-5 (default: gpt-4o-mini)
• tokenCount.displayOnRightSide: show status bar item on the right (default: false)
• tokenCount.showForUntitled: include untitled editors (default: true)
• tokenCount.debounceMs: debounce for normal files in ms (default: 120)
• tokenCount.largeFileDebounceMs: debounce for large files in ms (default: 450)
• tokenCount.largeFileCharThreshold: character threshold for large-file debounce (default: 60000)

Development

npm install
npm run compile
npm run lint
npm test

Run F5 in VS Code to launch an Extension Development Host.

Packaging and Publishing

npm run package:vsix
npm run package:pre-release
npm run publish:vsce
npm run publish:patch
npm run publish:minor
npm run publish:major