Find tools › Open-source projects › Spend less on AI

caveman

Open-source project · by JuliusBrussee

An add-on rule that makes AI coding helpers reply in very short sentences, so each answer costs less.

No codingVery popular

What it costs

Free and open source; you may pay for the AI model it uses.

Use it to

  • Cut the length of your coding agent's replies
  • Shrink long logs and test output before the AI reads them
  • Stretch a usage limit further during long coding sessions

How to start

  1. Install Node.js so the npx command works.

  2. In a terminal, run npx skills add JuliusBrussee/caveman -g.

  3. Use your coding agent as usual; its replies become much shorter.

  4. Compare your usage before and after, since short Q&A chats can cost more.

Good to know. The repo description's '65% of tokens' and the savings figures are the project's claims; its own README says it can cost more than it saves on short Q&A (the rule file adds ~1,000 input tokens per call) and does not help if you are billed per request. The skill is MIT, but the proxy runtime is under BSL-1.1 (not fully open source). Measure on your own workload.

Before you install

  • Checked by a personWe read what it does before listing it.
  • License: Apache-2.0You are allowed to use it under these terms.
  • Last updated 2026-11-18

More detail

Caveman is a skill (a rule file) that makes AI coding agents answer in very short 'caveman' sentences, plus an optional local proxy and middleware that shrink logs, test output and other text before the agent reads it.

AI usage is billed per token, and agents often write long explanations and read large tool outputs. Caveman targets both sides to lower cost and reading time, while leaving code, commands and error messages unchanged.

Similar tools

Descriptions are written in plain English from each project's own documentation and refreshed by our agent; this one was last checked 2026-09-29. Spotted a mistake? · Open in Find tools · © 2026 Which AI Map