Token Counter

Live Token Counter showing the token count for each message plus Claude and gpt 5 usage limits, in plain view before you hit them.

Token Counter is a lightweight extension that lives inside chatgpt.com and claude.ai, showing what every message really costs. The quota on your plan is real, yet neither page displays it, so people end up searching claude.ai/usage mid task and guessing at the remaining room. The gap between what you see and what you spend is exactly the problem this extension closes. Guessing is not a strategy. Guesswork it replaces: • Real message cost • What threads re-send • What features add • Where the quota went Token Counter keeps three numbers in view: 💠 Draft prices your typing live 💠 Chat totals the whole conversation 💠 Overhead covers prompt, features and memory The whole conversation is re-sent with every turn, so Chat climbs far faster than the words on screen. Put together, the three explain why a short reply can hold a heavy token count and why ChatGPT quotas and Claude usage limits arrive sooner than anyone expects. Token Counter turns that mystery into plain arithmetic. Each number updates independently, so the culprit is always obvious. What Token Counter adds: ➤ Counts while you type ➤ Whole chat in tokens ➤ Shows invisible overhead ➤ Local token counting Everything runs on your device with the open source openai tokenizer as the engine, so each count token check lands in milliseconds and nothing is uploaded anywhere. Modern tokenizers are quick, and you will never notice the work happening. Token Counter never asks for an API key and never touches your account. Speed was a design goal, not an afterthought. What it asks of you: - Zero setup - No login - One click - Fully local - No ads How the meter reads the page: 1. Token Counter loads on both sites 2. A local tokenizer prices drafts 3. Chat updates the token count 4. No anthropic token counter needed Privacy here is not a promise bolted on afterwards, it is the architecture. The measuring happens where the writing happens, so there is nothing to send, nothing to store and nothing to leak. Even a quick check of how many tokens is 100 words never leaves your machine. The safest data is the data that never travels, and this design makes travel impossible. In practice: ▸ No trace after uninstall ▸ No analytics, no trackers ▸ Nothing collected, ever Maybe you already paste prompts into an openai token counter, or track GPT 5 usage limits by hand, and know how calming a visible number is. A generic chatgpt calculator still stops at the prompt you typed and ignores what the thread drags along. Here one extension does both jobs: a claude token counter on this tab, a GPT meter on the next, keeping your claude usage in plain sight throughout. Who leans on the meter hardest, hour after hour: 1️⃣ Developers pasting long files daily 2️⃣ Writers grinding through all-day sessions 3️⃣ Analysts running an ai token counter 4️⃣ Pro, Max and Plus subscribers 5️⃣ Privacy minded users going local Two levers control almost everything, and Token Counter puts both on display. Long chats are the expensive part, so a fresh thread resets the biggest number in one move. Unused features quietly add tokens until you switch them off. A quota you can see is a quota you can steer, which is the whole point of an llm token counter that lives where you type and watches your claude limit for you. ❓ How many tokens is 100 words? 💡 Roughly 130, though simple wording lands lower and technical wording higher, which is why the Draft meter reads your actual text instead of estimating from length. ❓ Is this the token counter openai users already know? 💡 It plays that role on chatgpt.com and covers Claude too, since the same proven gpt tokenizer powers the math in both places. ❓ Will the anthropic token count match the official numbers? 💡 Closely. The claude tokenizer inside mirrors how Claude splits text, so the local estimate tracks what the API reports back. ❓ Where do claude ai usage limits actually appear? 💡 Only after you hit them. The meter counts tokens continuously, so the warning lands long before the lockout message does. No countdown drama, just an early heads-up. Install Token Counter today and turn an invisible quota into numbers you control. One click adds a live token calculator to every conversation and keeps your Claude usage limits where you can see them. The habit transfers in about a minute, and the meter fades into the background until you need it. The same install works as an open ai token counter on ChatGPT and gives Claude equal visibility.