RTK

RTK is a command-line tool that compresses terminal output to reduce token consumption for AI coding agents.

RTK is an open-source command-line tool that cleans up CLI output noise before it reaches AI coding agents. Written in Rust, it acts as a thin proxy layer that compresses verbose boilerplate and repetitive data while preserving test failures, errors, diffs, and stack traces. It works with Claude Code, Cursor, GitHub Copilot, Codex, Gemini CLI, Windsurf, Cline, OpenCode, and Aider. Developers install it via a one-liner, Homebrew, or pre-built binaries, and activate hooks to automatically rewrite shell commands. The tool is free and open source under the Apache 2.0 license, while RTK Pro offers team analytics, security guardrails, and governance features.

RTK at a glance

Best for: Software engineer
Runs on: macOS, Windows, Linux, CLI
Pricing: Free and open source; RTK Pro pricing available upon request for teams
Released: January 2026
Category: AI coding tool
Website: rtk-ai.app

Key features

Compresses command outputs automatically
Integrates via shell hooks and proxy layer
Preserves test failures and error messages
Tracks token savings with rtk gain
Provides team cost analytics and observability
Blocks secret leaks and stops dangerous actions
Runs locally as a single Rust binary

Use it for

Reducing token costs for AI coding agents
Cleaning up verbose test and build output
Preventing context window compaction
Enforcing security guardrails across developer teams

Companion to

Tools whose articles name RTK as a companion, each with its job for them.

Codex — RTK: Token savings

Alternatives to RTK

JFrog Boost — Lowers token usage for AI coding agents through input filtering.
jevgrep — Reduces coding agent costs by providing targeted, relevant repository context.
Reika — Manages small local models with careful context discipline and loop breakers.
CodeSight — Reduces AI token usage significantly and maps projects instantly.
Tokscale — Track and compare token usage across multiple AI coding agents.
Rubi — It cuts AI token waste by keeping sessions short and mapping project knowledge.
CodeGraph — It eliminates file crawling, reducing AI agent tool calls by 88% and cutting costs.
Genkit — Provides unified APIs and local developer tools for AI application workflows.
ccusage — Tracks token usage and costs locally across multiple coding agent CLIs.
Squeeze — Cuts token usage for coding agents by shrinking verbose tool output while retaining key data.
Command Code — Repairs open model tool calls and learns developer patterns automatically.
Tusk — Generates tests from production traffic to catch regressions in PRs.

All RTK alternatives