RTK: Rust-Based CLI Proxy to Reduce LLM Token Usage

Repositoryrtk-ai/rtkTool

RTK (Rust Token Killer) is a CLI proxy designed to filter and compress shell command outputs before they enter an LLM agent context. Distributed as a single, zero-dependency Rust binary, it aims to cut terminal output token consumption by 60 to 90 per cent.

Key points

  1. Filters output across 100+ commands with less than 10ms of execution overhead.
  2. Optimises outputs for tools such as git, ls, grep, docker, and test runners (npm, cargo, pytest, go test) by stripping boilerplate and reporting only failures or structural summaries.
  3. The 90 per cent saving applies specifically to bash output, which represents only a subset of total input tokens alongside system prompts and chat history.
  4. Implemented as a single binary in Rust under the Apache-2.0 licence with no runtime dependencies.

Why it matters

In agentic coding workflows, extensive terminal outputs rapidly consume context windows and increase API costs. Pre-filtering output directly in the shell eliminates verbosity, conserves context capacity, and lowers input token counts.

Analysis

Targeted at developers running terminal-based AI agents such as Claude Code, RTK addresses a direct operational bottleneck. With over 78,000 GitHub stars, Homebrew support, and multilingual documentation, the tool demonstrates high community traction and maturity. The documentation also remains clear that output compression does not equate to an equivalent reduction in overall LLM spend.

Mentioned links

Found in

Summary generated by AI (Gemini 3.8 Flash, 27 September 2026). It may contain errors – the original source is authoritative.

Related entries

Similar in content, found by the AI search.

  • Repository:decolua/9router

    9Router: Local Proxy and Token Saver for AI Coding Assistants

    9Router is an open-source AI routing proxy that connects coding tools such as Claude Code, Cursor, and Cline to more than 40 AI providers. It compresses tool results to save 20–40% in token usage and provides automated fallbacks across subscriptions and free models to prevent rate-limit interruptions.

    27.1kstarsJavaScript

    AI & LLMs· Tool

  • Repository:JuliusBrussee/caveman

    Caveman: Slashing Coding Agent Token Usage with Ultra-Concise Outputs

    Caveman is an open-source tool and prompt skill designed to strip conversational filler from AI coding assistants. By instructing models to communicate in an ultra-terse 'caveman' style, the project cuts output token counts by around 65% while keeping actual code, commands, and technical fixes intact.

    103.7kstarsGo

    AI & LLMs· Tool

  • Repository:headroomlabs-ai/headroom

    Headroom: Local Context Compression Layer for AI Agents and RAG

    Headroom is an open-source tool designed to compress tool outputs, logs, files, and RAG chunks before sending them to an LLM. It operates entirely locally, cutting token usage by 20% for coding agents and up to 95% for JSON data while preserving critical details.

    69kstarsPython

    AI & LLMs· Tool

  • Repository:getagentseal/codeburn

    CodeBurn: Local Token Usage and Cost Tracking for AI Coding Tools

    CodeBurn is an open-source, locally run utility that monitors token consumption and spending across AI coding assistants. It aggregates metrics across more than 37 tools and agents, breaking down expenditure by model, project, and task.

    10.8kstarsTypeScript

    AI & LLMs· Tool

  • Repository:troopdegen/mucho-texto

    mucho-texto: Structured Output Compression for AI Coding Agents

    mucho-texto is a skill designed for AI assistants like Claude Code, Cursor, and AGENTS.md workflows. It prevents agents from outputting verbose reports by categorising finished work into four distinct buckets and requiring an uncompressed, durable artifact before summarising.

    Shell

    AI & LLMs· Tool

  • Article:Alex Rios

    Token Savings Through Terse Agent Language

    Alex Rios discusses methods for cutting costs in AI coding tools like Claude Code and Cursor. One shared repository claims a 75 percent reduction by forcing assistants to talk in an extremely compressed style without pleasantries or articles.

    AI & LLMs· Opinion

Back to the knowledge base

Let's talk about your project

Get in touch

Locations

  • Mattersburg
    Johann Nepomuk Bergerstraße 7/2/14
    7210 Mattersburg, Austria
  • Vienna
    Ungargasse 64-66/3/404
    1030 Wien, Austria

Parts of this content were created with the assistance of AI.