← All edge tools
Token counter
Token counter — see how language models read your text
Count tokens for OpenAI, Llama, Gemma, Mistral and Qwen tokenizers and see every token highlighted. Useful for prompt budgets, API costs and context limits.
Stays on your computer — your text never goes to the internet
The AI model downloads to your computer once on first use, then everything is processed by your browser. Nothing is uploaded to a server — not even to Hypervision. No account, no storage, no tracking.