Token Counter for LLMs

Paste text, pick a model, see the count and where every boundary falls. Exact where the tokenizer is published, openly estimated where it is not, and never the two dressed the same. Everything runs in your browser, so your text never leaves the device.