▄ ▄▀▀ █▀▀▀ ▀█▀▀ ▀█▀▀ █▀▄▀█ █ █ █ █ ▀ █ ▀ ▀▀▀ ▀ ▀ ▀
rtfm explains the next step, why it works, and exactly where the docs cover it. You write the code, and actually learn the thing.
$npm i -g rtfm-cli
Right now every keystroke in handleChange fires a request. You want to wait until typing pauses. Three steps:
Watch out: if the search runs in the effect without that cleanup, you'll still get one request per keystroke, just delayed.
Nice, the state split is exactly right. Two things worth a second look:
The fix is yours. Cmd+click a path to jump there.
AI that writes code for you is great when you just need something built. rtfm is for when you want to learn: a new language or framework, a course project, or a codebase you want to understand.
One to three things you can do in the next ten minutes, explained clearly enough to write yourself. Come back for the next ones.
Every API comes with a link to the right section of the official docs, or the Stack Overflow answer you needed, so you learn where to look next time.
Stuck? Say so. Each hint gets more specific: the concept, then the API, then exactly what the line needs to do.
Press tab. It reads your git diff and points to the lines worth a second look, and why. The fix stays yours.
It can read your project and search the web, but it has no tools to change your files. Your code is your code.
Run it in the terminal of VS Code or any editor you like, and cmd+click a file:line to jump straight there.
$ npm i -g rtfm-cliOr brew install varunbln/tap/rtfm
$ cd my-app && rtfmA short walkthrough explains the two modes.
› /connectClaude, GPT, Gemini, DeepSeek, OpenRouter, or a local model via Ollama.
Free, open source, and yours to run with any model.
$npm i -g rtfm-cli