RTK Output Visualizer
See how rtk-style filtering shrinks command output before an AI agent reads it.
About rtk
rtk is open source (Apache 2.0). This page is not part of rtk and is not made by its authors.
Questions
What is rtk?
rtk (Rust Token Killer) is an open-source command-line proxy. It runs commands like git, ls and cargo test for an AI coding agent and filters their output first, so the agent reads fewer tokens.
Is this the same as running rtk?
No. These filters follow the four strategies rtk documents (filter noise, group, truncate, remove duplicates) for the commands it lists, but they are written for this page. Real rtk output can differ. To use rtk itself, install it from its GitHub page.
How is the command detected?
From the output itself: "On branch" means git status, "commit" with "Author:" means git log, "diff --git" means git diff, and so on. Anything else gets the general filter, which removes blank lines and folds repeated lines into one with a count.
Which tokenizer do the counts use?
OpenAI's o200k_base, used by GPT-4o. Claude and other models use their own tokenizers, so real counts differ, but the percentage saved is usually close.
Is my output uploaded?
No. Filtering and counting run in your browser. The tokenizer data downloads once from this site.
Related tools
- Token CounterCount tokens and see how text is split, with the real GPT-4o and GPT-4 tokenizers.
- LLM Sampling ExplorerSee what temperature, top-k, top-p and min-p do to a model's next-token choices.
- Whitespace CleanerRemove trailing spaces, mixed line endings and extra blank lines.
- Text DiffCompare two texts line by line or word by word, highlighted live.