Neura Market
caveman logo

caveman

Free

🪨 why use many token when few token do trick — Claude Code skill that cuts 65% of tokens by talking like caveman

Model APIsFreeFree tier
Inputs: textOutputs: text, code
Type
Open Source
Visit Website GitHub (52.8k stars)

About caveman

Caveman is an open-source, token-efficient stack for agent-native development that reduces token usage by 65% on average by transforming verbose instructions into concise, caveman-style commands. It integrates as a Claude Code skill, an npm CLI tool (Caveman Code), a browser extension (for ChatGPT, Claude, and Gemini), and a separate Cavemem package. The compression engine is byte-safe and includes a local compression workbench for testing and visualization. Caveman is free to self-host with your own API keys (MIT license) and also offers a hosted cloud gateway with verified savings, eval-gated rollout, and receipt verification for paid plans. Trusted by 10,000,000+ professionals and used by companies like OpenAI, Microsoft, Vercel, and Cloudflare.

Key Features

Token compression that reduces token usage by 65% on average
Claude Code skill integration for direct CLI usage
Caveman Code npm package for programmatic compression
Browser extension supporting ChatGPT, Claude, and Gemini
Byte-safe compression gateway ensuring data integrity
Compression workbench with local demo and token visualization
Open source (MIT license) and free to self-host with own API keys
Multi-platform support: Claude Code, npm CLI, browser, Cavemem
Caveman Labs research division with fine-tuned models (CaveGemma)
Trusted by 10M+ professionals and major tech companies

Pros & Cons

Pros
  • Cuts token usage by 65% on average, reducing AI API costs
  • Open source (MIT) and free to self-host with own API keys
  • Multiple integration methods: Claude Code skill, npm CLI, browser extension
  • Byte-safe compression gateway ensures data integrity
  • Compression workbench allows testing and visualization of token savings
  • Trusted by 10M+ professionals and companies like OpenAI, Microsoft, Vercel, Cloudflare
  • Supports multiple platforms: Claude, ChatGPT, Gemini, and npm
Cons
  • Token savings are inferred and not verified without the hosted cloud gateway
  • Cloud features (dashboard, eval-gated rollout, receipt verification) require waitlist and paid plans
  • Compression may reduce readability and require adoption of 'caveman' style prompts
  • Only applies to text prompts, not other data types or multimodal inputs
  • Limited to supported agents and platforms (Claude Code, npm, browser extensions)

Best For

Reducing token costs in AI agent development workflowsCompressing verbose prompts for Claude, ChatGPT, and GeminiOptimizing token usage in CI/CD pipelines with agent-native toolsDecreasing API latency and spending for teams using large language modelsTesting token savings interactively with the compression workbench

Alternatives to caveman

AutoGPT

AutoGPT is the vision of accessible AI for everyone, to use and to build on. Our mission is to provide the tools, so that you can focus on what matters.

aider

aider is AI pair programming in your terminal

axolotl

Go ahead and axolotl questions

Chroma

Open-source embedding database

awesome-claude-code

A curated list of awesome skills, hooks, slash-commands, agent orchestrators, applications, and plugins for Claude Code by Anthropic

agents-course

This repository contains the Hugging Face Agents Course.

FAQ

What is Caveman?
Caveman is an open-source token-efficient stack for agent-native development that compresses verbose instructions into concise, caveman-style prompts, cutting token usage by 65% on average.
How do I install Caveman?
You can install via curl script (curl -fsSL https://raw.githubusercontent.com/JuliusBrussee/caveman/main/install.sh | bash), npm install @juliusbrussee/caveman-code, or as a browser extension for ChatGPT, Claude, and Gemini.
Is Caveman free?
Yes, the open toolkit is free forever and self-hosted with your own API keys. Optional hosted cloud features have paid plans (Indie $19/mo, Team $299/mo, Enterprise custom).
How much can I save on tokens?
On average, Caveman cuts 65% of tokens. The compression workbench on the website allows you to test savings with your own prompts, with up to 69% savings demonstrated in examples.
What platforms does Caveman support?
It supports Claude Code as a skill, npm CLI (Caveman Code), browser extensions for ChatGPT, Claude, and Gemini, and a separate Cavemem package. A hosted cloud gateway is also available via waitlist.