workflowstacks
claude-skill

Rtk

CLI proxy that reduces LLM token consumption by 60-90% on common dev commands. Single Rust binary, zero dependencies
76,019 stars4,784 forksRustUpdated 8/13/2026100% free · open source
What it does

RTK is a CLI proxy that reduces Large Language Model (LLM) token consumption by 60-90% on common development commands.

Install / run
cargo install rtk
When to use it
  • When developing applications that rely heavily on LLMs and token usage needs to be optimized
  • When using LLMs for tasks like code completion, text summarization, or natural language processing
  • When trying to reduce costs associated with LLM token consumption in development workflows
Quick start
  1. 1Run `rtk init` to initialize the RTK configuration
  2. 2Configure your LLM API endpoint and credentials in the `rtk.yaml` file
  3. 3Use `rtk` as a proxy for your LLM commands, for example `rtk my-llm-command`
  4. 4Verify token consumption reduction by comparing the output of `rtk my-llm-command` with the original command
  5. 5Tune the `rtk.yaml` configuration to optimize token usage for your specific use case
Ready-to-paste prompt
rtk my-llm-model --prompt 'Summarize the following text: ...'
Heads up: RTK requires a valid LLM API endpoint and credentials configured in the `rtk.yaml` file to function correctly
Saves to your device

Topics

agentic-coding
ai-coding
anthropic
claude-code
cli
command-line-tool
cost-reduction
developer-tools
llm
open-source
productivity
rust
token-optimization
How Rtk works
Codeflow
Free to inspect

Rtk is a very large Rust project (~73k lines across 154 code files, plus 15 test files). Setup is light: installs like a normal app. Reading the code is optional. Last commit this month, Apache-2.0 license, has a test suite.

Size
Very large codebase
~73k lines · 154 code files · days to read — use, don't read
Setup
One-command install
Installs like a normal app. Reading the code is optional.
Runs on
Rust
No API keys detected
Rust 94%Shell 4%TypeScript 1%Python 1%
What happens, step by step

Filters command outputs for AI tools

You give
shell commands
1Intercept
Catches shell commands
src/main.rs
2Compress
Reduces command output
3Filter
Removes unnecessary data
4Report
Shows savings dashboard
src/main.rs
You get
compressed outputs
Where to start reading
  1. 1
    README.md
    Start here — what it does and how to install it
  2. 2
    CLAUDE.md
    The instructions the AI actually follows
  3. 3
    src/main.rs
    Where the program starts running
  4. 4
    Cargo.toml
    Dependencies and the commands it exposes
What's in each folder
src/Core code — the actual logic203 files
docs/Documentation19 files
hooks/Hooks — scripts that auto-run on events30 files
scripts/Helper scripts20 files
openclaw/Folder4 files
Formula/Folder1 files
tests/Tests — proof it works60 files
.claude/Editor / agent settings40 files
READMEHas testsDocumentedCI checksApache-2.0 licenseUpdated this month
Quick Actions
Details
Creator
rtk-ai
Language
Rust
Category
claude-skill
Published
1/22/2026

Are you the creator of this tool? Claim your listing → and earn 85% of every sale.