Click to select a file or drag & drop
Supported formats: .txt, .md, .json, .js, .py, .html, .css, .xml, .csvToken Analysis
About Token Calculation
OpenAI Models
Uses js-tiktoken library for accurate BPE tokenization of GPT models. Tokens are the basic units of text processing.
Grok Models
Uses approximate tokenization similar to GPT models. Grok typically uses ~3.9 characters per token with a hybrid calculation combining character-based (3.9 chars/token) and word-based (1.35 tokens/word) estimates for improved accuracy.
Phi Models
Uses approximate tokenization similar to GPT-style BPE. Phi estimates use a hybrid approach: ~4 characters per token and ~1.3 tokens per word, averaged for stability.
DeepSeek Models
Uses approximate tokenization similar to GPT models. DeepSeek typically uses ~4 characters per token with a hybrid calculation combining character-based (4 chars/token) and word-based (1.3 tokens/word) estimates, then averages both methods.
Claude Models
Uses approximate tokenization based on character count. Claude typically uses ~4 characters per token.
Privacy
All processing happens in your browser. No text is sent to external servers.
- All text processing happens locally in your browser
- No data is sent to external servers
- No tracking or analytics
- Your text never leaves your device
- No cookies or local storage used for tracking