CONTACT
Questions? Bug reports? Enterprise deals? We read everything.
SEND MESSAGE
No ticket system. No bot. Direct to inbox.
QUICK ANSWERS
Is CuToken free to use?
Yes.No SignUp, No cost, no license keys, no phone home.
How accurate is the token counting?
Approximate (~3.8-4.0 chars/token). For exact counts, use the Python version with tiktoken or a PHP tiktoken extension. The optimizer works on relative savings, so approximations are consistent.
Can it break my prompt meaning?
Safe mode: never. Balanced/Aggressive: guarded by grammar checks (dangling prepositions, doubled determiners, modal+to, balanced quotes). Confidence score indicates risk. Review output for aggressive mode.
Does it work with non-English prompts?
Optimized for English. Filler/hedge lists are English-centric. Phrase compression may work for similar structures. Non-Latin scripts will have different token ratios.
What's the max prompt length?
No hard limit in the optimizer. Practical limit is your model's context window minus completion budget. The web demo caps at ~4000 chars for UX.
Can I self-host the API?
No — CuToken isn't an open-source or self-hostable platform. That said, no prompt data is stored or logged when you use it.