llm-tokencheck: standalone CLI to estimate cost across Anthropic/OpenAI/Google before sending (no llm dependency)

[ilan-lachkar]

I kept overrunning my API budget by guessing token counts before sending big prompts, so I built llm-tokencheck: a small free CLI that estimates tokens + USD cost for a file or string across current Anthropic, OpenAI and Google pricing, before you make the call.

npm install -g llm-tokencheck
tokencheck myfile.txt

It's a standalone tool (no dependency on llm itself), so it also works if you're calling APIs directly from your own code/scripts and just want a quick multi-provider sanity check before sending.

I noticed llm-confirm-tokens and llm-cost already posted here — good tools, but they're llm plugins (pre-flight gate / post-hoc log report over llm's own logs). This one is provider-agnostic and doesn't require llm to be installed, aimed more at "should I even send this, and to which model" before any code is written.

• GitHub: https://github.com/ilan-lachkar/llm-tokencheck
• npm: https://www.npmjs.com/package/llm-tokencheck

Free and open source. There's an optional $9 Pro license for batch folder processing + CSV export — disclosing it since I'm the author.

Feedback on pricing accuracy for newer models very welcome.