# rtk-ai/rtk

CLI proxy that reduces LLM token consumption by 60-90% on common dev commands. Single Rust binary, zero dependencies

Repository: https://github.com/rtk-ai/rtk
Canonical: https://ross.abutalabs.com/products/rtk
Homepage: https://www.rtk-ai.app
Language: Rust
License: Apache-2.0
License Family: permissive
Topics: agentic-coding, ai-coding, anthropic, claude-code, cli, command-line-tool, cost-reduction, developer-tools, llm, open-source, productivity, rust, token-optimization
Last push: 2026-08-26T23:36:54+00:00

## Health v2 (maintenance only)
Score: 78/100 (v2, computed 2026-09-03T02:39:23.370411+00:00)
- activity 99, release rhythm 87, longevity 15
- inputs: {"age_days": 223, "days_push": 7, "days_rel": 7, "gap_med": 0.0, "n_releases_24m": 81}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 77533, forks 4875 (observed 2026-08-28T04:12:21.641782+00:00)

## What it is
RTK (Rust Token Killer) is a CLI proxy that intercepts shell commands and compresses their output before it reaches an LLM agent's context window, cutting bash output tokens by 60-90%. It ships as a single zero-dependency Rust binary supporting 100+ common commands like git, cargo test, pytest, grep, and ls.

## Use cases
- reduce token costs when using AI coding agents like Claude Code or Cursor
- prevent context window overflow during long agent coding sessions
- compress git diff and git status output for LLM consumption
- trim test runner output to failures only for AI agents
- lower API bills on pay-per-token coding assistants
- make CLI output more compact for human reading too

## When to choose
- you use AI coding agents (Claude Code, Aider, Cursor, Gemini CLI) that read lots of shell output
- you hit context limits or rate limits during agent sessions
- you pay per token and want to cut CLI noise from your bill
- you want a fast, zero-dependency single-binary tool with <10ms overhead

## When to avoid
- you need full, unabridged command output for debugging or auditing
- your workflow depends on parsing exact raw output formats from commands
- you don't use LLM agents and aren't bothered by verbose CLI output
- you need guaranteed identical output semantics from wrapped commands

## Facets
- artifact type: cli-tool
- maturity: active
- function: cli, developer-tools, llm-inference, compression, middleware
- domain: developer-tools, large-language-models
- platform: cli, cross-platform, windows
- tags: token-optimization, agentic-coding, claude-code, cost-reduction, rust-binary, cli-proxy, command-line, ai-agents, macos, linux

## Member repositories
- rtk-ai/rtk (main) score 78

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:12:21.641782+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T16:14:20.376451+00:00, confidence not recorded.
  - readme: https://github.com/rtk-ai/rtk (fetched 2026-08-28T04:12:21.641782+00:00, sha 088c508d172e)
  - homepage: https://www.rtk-ai.app (fetched 2026-08-28T17:42:51.511470+00:00, sha 5615da3daf6e)
  - site_page: https://www.rtk-ai.app/guide/ (fetched 2026-08-28T17:42:51.520248+00:00, sha af37812a8752)
  - site_page: https://www.rtk-ai.app/guide/getting-started/installation/ (fetched 2026-08-28T17:42:51.521991+00:00, sha 34b4331aceee)
  - site_page: https://www.rtk-ai.app/guide/getting-started/quick-start/ (fetched 2026-08-28T17:42:51.523891+00:00, sha b379610d89d5)
- Data as of 2026-08-30T08:39:29.467469+00:00.