Free tools
Count characters, CJK glyphs and tokens before you spend them
Paste text or drop a file. You get character, CJK, word and byte counts plus an estimated token cost and how many 700-character chunks it would become.
- Files never leave the device
- No account, no tracking pixels
- Not used for training
- Works offline once cached
Read more
- CJK glyphs and Latin words are counted separately, because their token ratios differ by roughly 2.7x.
- The token figure is an estimate (about 4 characters per token for Latin, 1.5 for CJK) — real tokenizers differ per model.
- The chunk count uses the same 700-character / 15% overlap setting as the search tool.
- Everything runs in the page; no request leaves the tab.