Caveman Code
Caveman Code is a terminal coding agent designed to minimize token usage through aggressive, multi-layer context compression. It features an autonomous goal loop, persistent semantic memory, and a plan mode for read-only exploration.
Project status is frozen as of August 2026; development has shifted to the 'caveman' proxy engine, which provides similar token-saving capabilities to existing agents.
As of
- 4-layer token compression (prompts, tool output, file reads)
- Autonomous goal loop with shadow-git checkpoints
- Persistent semantic memory via cavemem
- Support for 20+ AI providers via OAuth and API keys
- Architect/editor model split for cost-efficient workflows
Developers looking to reduce AI coding agent costs by minimizing redundant token consumption in large codebases.
Users requiring active software support or updates, as the project is in a frozen state.
Open-source (MIT License); free to install and use.
An efficient, open-source coding agent that succeeds at significant token reduction, though it is no longer under active development and acts primarily as a historical milestone in the creator's broader AI efficiency research.
Is Caveman Code free?
Yes - Caveman Code is Free. Open-source (MIT License); free to install and use.
Is Caveman Code open source?
Yes - Caveman Code is open source. Open-source (MIT License); free to install and use.
Who is Caveman Code best for?
Developers looking to reduce AI coding agent costs by minimizing redundant token consumption in large codebases.
Who is Caveman Code not ideal for?
Users requiring active software support or updates, as the project is in a frozen state.
What are the best Caveman Code alternatives?
The closest Caveman Code alternatives on ai.dosa.dev are Claude Code, Codex CLI, Amp - all listed under Terminal & CLI Agents.
Who makes Caveman Code?
Caveman Code is developed by Julius Brussee. It is listed in the Terminal & CLI Agents category on ai.dosa.dev.
Favorite this tool to revisit it later, or Zap it to contribute to the public vote count.
Content on this page is AI-generated. Please verify details with the vendor's website for accuracy.