Token counter
Paste your text to see roughly how many tokens it uses before you send it to an AI chat. Counting happens in your browser — nothing is uploaded.
Paste some text to see the ratio of characters to tokens.
What is a token?
A token is the unit an AI model actually reads. It is usually a fragment of a word rather than a whole word or a single letter. Common words such as the or hello are one token each; longer or unusual words are broken into several pieces.
That is why character counts mislead. English prose runs at roughly four characters per token, code is denser at about two to three, and Chinese or Japanese can approach one token per character. The same 2,000 characters might be 500 tokens of English or close to 2,000 tokens of Chinese — so a limit that comfortably fits one will be rejected for the other.
Why AI chats count tokens
A model's context window — how much it can hold in mind at once — is measured in tokens, so that is what services meter and bill. When a paste is refused or silently truncated, it is almost always a token limit you have hit, not a character limit. Counting tokens before you paste tells you whether the text will fit.
How exact is this?
It is a close estimate, not an exact count, and we would rather say so than overclaim. ChatGPT, Claude and Gemini each use a different tokenizer, and Claude's is not published, so no single number can be exact for all three at once. This counter is tuned to sit slightly high rather than slightly low, so text that fits according to the estimate will fit in practice. For everyday English prose it is typically within a few per cent.
Nothing leaves your browser
The counting runs in JavaScript on this page. Your text is never uploaded, logged or stored, and once the page has loaded it keeps working with no connection at all. That is the same promise as the rest of SplitMyText — you can check it in your browser's network tools.
When your text is too long
Split it into numbered parts and send them in order, asking the model to wait until every part has arrived before it replies. The free web tool does this in your browser, and the iPhone, iPad and Mac app can split by token count directly, so each part lands inside the limit you set.
For a walkthrough of the whole approach, see pasting long text into AI chats. For the limits on other platforms, see the character limits table.