Runs 100% in your browser
Your data never leaves your device. zero uploads.
Estimate the token length of AI prompts and inspect how the selected tokenizer affects the count.
List prices as of June 2026, per 1M tokens. Verify current rates with each provider.
| Model | Tokens | Total cost |
|---|---|---|
GPT-5 OpenAIexact | 119 | $0.00515 |
GPT-5 mini OpenAIexact | 119 | $0.00103 |
GPT-4.1 OpenAIexact | 119 | $0.00424 |
GPT-4o OpenAIexact | 119 | $0.00530 |
GPT-4o mini OpenAIexact | 119 | $0.00032 |
o3 / o4-mini OpenAIexact | 119 | $0.00424 |
GPT-4 Turbo OpenAIexact | 119 | $0.0162 |
GPT-3.5 Turbo OpenAIexact | 119 | $0.00081 |
Claude Opus 4 Anthropicest | 133 | $0.0132 |
Claude Sonnet 4 Anthropicest | 133 | $0.00790 |
Claude Haiku 4.5 Anthropicest | 133 | $0.00263 |
Gemini 3 Pro Googleest | 119 | $0.00624 |
Gemini 3 Flash Googleest | 119 | $0.00156 |
Gemini 2.0 Flash Googleest | 119 | $0.00021 |
Llama 3.1 405B Metaest | 126 | $0.00056 |
DeepSeek V3 DeepSeekest | 129 | $0.00058 |
Grok 4 xAIest | 123 | $0.00787 |
Mistral Large Mistralest | 129 | $0.00326 |
Check a draft prompt, context excerpt, or generated response before combining it with other material. Compare the available model selections and leave room for the model's response when planning context usage. The calculation runs on the text in the workspace.
Count the instructions and context separately, then measure their combined text to assess the size of a proposed prompt.
The displayed count or cost estimate is not an API billing quote. Verify model-specific limits, overhead, and current pricing with your provider.
Paste the text you want to measure. Choose a model or tokenizer from the available options. Review the token count and any displayed estimate before planning your prompt.
The displayed count or cost estimate is not an API billing quote. Verify model-specific limits, overhead, and current pricing with your provider.
No registration, no tracking of your content, no server uploads. Don't just take our word for it:
Your files and text are processed entirely on your device. Nothing is ever sent to us.
Once loaded, this tool keeps working with your Wi-Fi turned off. Try it.
Open your browser's DevTools → Network tab. You'll see zero requests carrying your data.
More utilities for your next task.
Count tokens in text with local tokenizers and compare estimates across popular AI models like GPT, Claude, and Llama. Free and private, no API key needed.
Encode JWT header and payload data and sign test tokens for development and debugging. Free online JWT encoder that runs locally, keys never leave your device.
Decode a JWT header and payload to inspect its claims, expiry, and issued-at times. Free online JWT decoder; tokens stay in your browser and are never sent.