Media
2 items
Description
Live token count while typing in ChatGPT (GPT-5.5). 100% local tokenization — your prompts never leave your browser. See your prompt's exact token count as you type in ChatGPT — no more guessing whether you're about to blow past the context window. Works with the latest models Built for GPT-5.5 Instant, GPT-5.5 Thinking, and GPT-5.5 Pro (and falls back correctly for GPT-4o, o1, o3, and earlier GPT-5 models — anything on OpenAI's o200k_base tokenizer). Exact counts, not estimates Most token counters guess using a "~4 characters per token" rule of thumb. This extension runs the real o200k_base byte-pair encoding algorithm — the same tokenizer OpenAI's own tiktoken library uses — so the number you see is the number that actually counts against your context window. See how much room you have left Optionally show what percentage of the model's context window your current prompt is using, so you know at a glance whether you're at 2% or 95%. Stays out of your way A small, draggable pill appears near the compose box. Move it anywhere, collapse it, or turn it off entirely from the toolbar popup. Privacy first Your prompt text is tokenized entirely inside your own browser. Nothing you type is ever sent to us or to any third party. The only network request this extension ever makes is a one-time download of OpenAI's own public tokenizer data file (the same file their official tiktoken library downloads) — after that, everything works fully offline. Settings - Show/hide the counter - Show % of context window used - Pick your plan and model (GPT-5.5 Instant Free/Plus/Pro, GPT-5.5 Thinking, GPT-5.5 Pro) to match ChatGPT's real in-app limits, or set a custom window size - Optional cost estimate using a price you enter yourself (we don't hardcode pricing, since it changes) - Reset the pill's position Feedback and bug reports: abdel.zakaria@gmail.com