100% free no uploads no sign-up

token counter

Paste a prompt, see the tokens. Exact counts for OpenAI models, honest estimates for Claude and Gemini — and your text never leaves your browser.

OpenAI o200k_base tokens (exact)loading…
OpenAI cl100k_base—
Claude (estimate)~22
Gemini (estimate)~19
characters77
words15

OpenAI counts are exact (same tokenizer OpenAI uses). Claude ≈ characters ÷ 3.5 (Anthropic's rule of thumb); Gemini ≈ characters ÷ 4 (Google's rule of thumb).

Why tokens matter

AI models don't read words — they read tokens. Every model has a context window (how many tokens it can consider at once) and API usage is billed per million tokens, separately for input and output. Knowing the token count of a long prompt, a document or a system message tells you whether it will fit and roughly what it will cost before you send it.

How the counts are made

Rules of thumb work best for English prose. Code, numbers, URLs and non-Latin scripts like Hindi usually take more tokens than the estimate suggests — the coloured view above shows exactly where OpenAI's tokenizer splits your text, which is a good guide to what's expensive.

Tips to use fewer tokens

Comparing AI subscriptions instead of API usage? See AI plan prices compared. Need a plain word count? Use the word counter.

Frequently asked questions

What is a token?

A token is the chunk of text an AI model reads and writes — often a whole short word, part of a longer word, a space plus a word, or a punctuation mark. Models have limits and prices measured in tokens, not words.

How many words is 1,000 tokens?

For typical English, roughly 700–800 words. Google's docs put 100 Gemini tokens at about 60–80 English words. Hindi, Tamil and other non-Latin scripts usually need noticeably more tokens per word.

Are the Claude and Gemini counts exact?

No — they're estimates using the rules of thumb Anthropic (≈3.5 characters per token) and Google (≈4 characters per token) publish. Their tokenizers aren't available to run in a browser. For exact counts, use each company's token-counting API.

Which encoding should I pick for OpenAI?

o200k_base is used by GPT-4o and newer OpenAI models. cl100k_base is used by GPT-4, GPT-3.5 Turbo and the text-embedding-3 models.

Is my text sent anywhere?

No. The tokenizer is downloaded once into your browser and all counting happens on your device.