Useful Projects
Open-source projects that cut token costs and add memory to your coding agents, whether the model runs locally or in the cloud.
headroomlabs-ai/headroom
78k
headroom
Compresses tool outputs, logs, files, and RAG chunks before they reach the model, cutting tokens by 20% for coding agents and 60-95% for JSON. Runs as a library, proxy, or MCP server.
rtk-ai/rtk
85k
rtk
A CLI proxy shipped as a single Rust binary with no dependencies. It cuts LLM token use by 60-90% on common dev commands.
cwinvestments/memstack
430
memstack
Adds persistent memory to Claude Code with a session diary and fact recall, plus 87 free skills and an MCP skill loader.