How to use
- Paste the draft, prompt, or document fragment.
- Set a token budget if you want a quick usage percentage.
- Use the CJK-adjusted estimate when the text mixes Japanese, Chinese, or Korean heavily.
Wave 40 breadth expansion
Estimate text size before it reaches a model or workflow limit
This page is intentionally rough. It is for planning and triage, not for billing-grade model accounting.
How to choose between nearby pages
Use Query String Parser or JSON Formatter when the question is structured payload inspection. Use this page when the question is plain-text length, rough prompt size, or editorial budgeting.
FAQ
What counts does this page show?
It shows characters, characters without spaces, words, lines, bytes, CJK character count, and two rough token estimates.
Why are there two token estimates?
One estimate is a general rough approximation and the other is adjusted to be less misleading for CJK-heavy text. Neither is model-specific.
Does the share URL include the text I pasted?
No. The share URL keeps only the optional token budget. The text stays in your browser.
When should I use this page instead of a model-specific tokenizer?
Use this page for quick planning, editorial checks, and rough budgeting. Use a model-specific tokenizer when billing or prompt-limit accuracy must match one exact model.