vincentkoc/tokenjuice

🧃 Token weight loss. Lean output compaction for terminal-heavy agent workflows. Works as a native CLI tool or as an extension to popular coding and agent frameworks.

What it solves

Terminal-based AI agents often struggle with "noisy" command output. When an agent runs commands like git status or docker build, the massive amount of text can clutter the AI's context window, leading to wasted tokens and inefficient performance. tokenjuice solves this by automatically shrinking that output into a concise, relevant summary.

How it works

tokenjuice acts as a deterministic filter. It observes the output of terminal commands after they execute and applies rule-driven "reducers" to compress the text. It keeps the original command semantics intact while ensuring only the essential information is sent back to the AI. The raw output is not lost; it remains accessible if you specifically request it or if you opt-in to artifact storage.

Who it’s for

It is designed for developers and power users who rely on terminal-heavy AI coding agents and want to optimize their workflows, reduce token consumption, and keep their AI interactions clean and focused.

Highlights

  • Deterministic output compaction using inspectable JSON rules rather than unpredictable AI logic.
  • Seamless integration with a wide variety of popular AI coding agents and CLI tools.
  • Preserves command semantics while reducing transcript waste.
  • Optional artifact storage allows for full recovery of raw output when needed.
  • Lightweight design that acts as a thin wrapper around existing agent workflows.

Related

  • Project
    rtk-ai/rtk

    A high-performance CLI proxy that compresses bash output for AI agents, reducing input tokens by up to 90% to lower costs and save context space.

  • Project
    zdk/lowfat

    lowfat is a lightweight, local‑only CLI that filters and compresses command‑line and file output before it reaches an LLM agent, offering three aggressiveness levels (lite/full/ultra) and plug‑in extensibility. It integrates with Claude Code, OpenCode, generic shells, and Pi agents, and provides local analytics (`lowfat stats`, `history`) to track token savings. Installation is via Cargo or Homebrew, and the tool is Apache‑2.0 licensed.

  • Project
    alexgreensh/token-optimizer

    Token Optimizer is a zero‑dependency Python tool that plugs into AI coding assistants (Claude Code, Copilot, etc.) to automatically compress command output, search results, JSON, file diffs and other “structural” data, saving tokens and money. It adds smart checkpoints so savings survive model compaction, scores context quality live, stores a local SQLite audit trail, and provides slash‑commands (e.g., `/token‑optimizer`, `/token‑coach`) for reports and dashboards. Installation is a one‑time plugin add, after which it runs silently in the background.

  • Project
    InterfaceX-co-jp/genshijin

    A token-compression framework for AI coding agents that reduces Japanese response redundancy to lower API costs and save context space.