Token Counter
Estimate tokens and rough GPT/Claude input cost in your browser.
Estimate tokens and rough API cost in your browser. Counts are approximations (not exact tiktoken for every model). Nothing is uploaded.
Estimate
Rough input cost (USD) at common list prices — verify with your provider:
About the Token Counter
AI models split text into tokens. APIs usually bill by token, and each model has a context window limit. This free token counter estimates tokens from your text in the browser and shows a rough input-cost sketch for common GPT and Claude list prices.
Counts are approximations (character/word heuristics), not an exact OpenAI tiktoken bill for every model. Use them to plan prompts — verify billing with your provider.
How to use
- Paste your prompt or document.
- Read characters, words, approximate tokens, and share of a 128k context.
- Check the rough USD input-cost estimates before calling a paid API.
Tips
- Long pasted web pages burn tokens fast — trim to what the model needs.
- Code and non-English text can tokenize differently; treat numbers as guidance.
- Pair with AI Commands when you want a clearer instruction before you send a large draft.
FAQ
Is this exact for GPT-4o? No. It is a practical estimate. For exact OpenAI counts use an official tokenizer in your own environment.
Does text leave my device? No. Calculation runs in your browser on this page.