A CLI proxy that compresses shell output before your coding agent reads it, so every command costs fewer tokens.
brew install rtkRTK sits between your coding agent and the shell. It runs common dev commands (git, ls, test runners, linters, package managers) and compresses the output before the agent sees it. Passing tests collapse to a count, file lists become a tree, and progress noise is dropped.
Install it for your OS:
brew install rtk # macOS / Linux
winget install rtk-ai.rtk # Windows
cargo install --git https://github.com/rtk-ai/rtk # from source
Then prefix commands with rtk, for example rtk git status or rtk npm test. Run rtk init -g to install the hook for your agent, so this happens without the prefix.
cargo install rtk.yamadashy · Tool / CLI
Packs a whole repository into one AI-friendly file, with token counts, so you can hand a codebase to any LLM.
NVIDIA NeMo · Tool / CLI
NVIDIA's Rust proxy that routes each LLM call to the cheapest model that can still do the job, keeping native OpenAI and Anthropic APIs.
ccusage · Tool / CLI
A CLI that reads your local Claude Code logs and shows token use and estimated cost by day, month and session.