Introducing the AI Token Counter
We built a browser-based token counter that uses the actual BPE tokenizer for OpenAI models — so you can see exactly what you're paying for before sending a single API request.
If you've ever been surprised by an OpenAI API bill, the cause is almost always the same: tokens accumulate faster than you expect, and you don't see the count until after you've paid.
We built the Token Counter to fix that feedback loop. It runs the actual js-tiktoken BPE tokenizer in the browser — the same tokenizer OpenAI uses — so the count you see is the count you'll be billed for.
What it does
Paste any text into the input and you instantly see:
- Exact token count for GPT-4o, GPT-4o mini, o1, and other OpenAI models
- Per-token visualization — each token is highlighted in a different color, so you can see where your prompt is "expensive"
- ChatML chat mode — wraps your text in the system/user/assistant template that OpenAI charges for in chat completions
- Rough estimates for Claude, Gemini, and DeepSeek (their tokenizers are proprietary, but we show honest approximations marked with
~)
Why it matters
A 500-token system prompt that runs 1 million times per month costs $75 at GPT-4o mini rates. If you can cut it to 200 tokens, that's $45/month saved — without changing a line of application code.
The Token Counter makes that optimization visible. Paste your system prompt, see the count, and start trimming.
It's fully client-side — your prompts never leave the browser.