paste text, get tokens. uses o200k tokenizer (tiktoken). runs locally.

context usage (128K window) 0%
output
0 tokens
0 chars
0 words
0 lines
note: uses OpenAI's o200k tokenizer (GPT-4o and newer). older OpenAI models use cl100k, and Claude, Gemini and the rest tokenize differently again, so for those the count is close rather than exact.