llm-tokencheck

Free token & USD cost estimator for Anthropic / OpenAI / Google API calls — paste text below, no install, no signup.

0 characters · 0 words · ~0 tokens (estimate)

Provider	Model	Input cost	Output cost*

*Output cost assumes a response of roughly the same token length.

Want this on the command line? Same estimate, works on any file or folder, no browser needed:

npm install -g llm-tokencheck
tokencheck myfile.txt

There's an optional $9 Pro license that unlocks batch processing of a whole folder + CSV export — useful if you're auditing costs across a codebase or a dataset.

Get the $9 Pro license → View source on GitHub

Already have existing API usage you suspect is more expensive than it needs to be (not just a one-off prompt)? llm-cost-audit ($29, one-time) manually reviews your usage export and reports concrete savings.

Launching your own dev tool and wondering where to actually post it? launchmap ($7) is a written report of 17+ distribution channels tested for real — what's blocked, what works.

How the estimate works

No external tokenizer dependency (keeps the CLI install at zero deps): the estimate blends a characters/4 approximation with a word-count based approximation, which stays within roughly 15% of real BPE tokenizers for typical English/French prose. Always shown as an estimate, never as an exact count — for exact counts, use the tokenizer your provider ships (e.g. tiktoken for OpenAI).