Caveman

All projects

Apps & Tools

Output compression for Claude Code and coding agents that reduces tokens by 65%.

byJulius Brussee

PlayableOpen source

Caveman media is blocked

Allow external media to connect to the provider and play this content.

Description

A developer tool and skill designed to compress LLM output for Claude Code and over 30 other agents. It removes filler tokens from model responses while keeping code, commands, and errors byte-for-byte exact, offering multiple intensity levels including a classical Chinese 'wenyan' mode.

Descriptions, tags, and model credits may be AI-generated or inferred from public sources and can be incomplete or wrong. Learn more.

Details

Why this is interesting: It addresses a major friction point in AI development—token costs and context limits—by providing a practical,…
Attributed to: @JuliusBrussee
Built with: K2.7 Code North Mini Code GPT Image 2 Redwood AI Codex Phi-3 Mini 4K
Links: caveman.so
View tweet
View GitHub repo
Tags: #compression #developer-tools #token-optimization #coding-agents

Discovered: X
First posted: X · Aug 9, 2026
Listed: Added 17d ago
Build: Strong
Models: Strong

Caveman

Description

Timeline

Caveman

Description

Timeline