npm package discovery and stats viewer.

Discover Tips

  • General search

    [free text search, go nuts!]

  • Package details

    pkg:[package-name]

  • User packages

    @[username]

Sponsor

Optimize Toolset

I’ve always been into building performant and accessible sites, but lately I’ve been taking it extremely seriously. So much so that I’ve been building a tool to help me optimize and monitor the sites that I build to make sure that I’m making an attempt to offer the best experience to those who visit them. If you’re into performant, accessible and SEO friendly sites, you might like it too! You can check it out at Optimize Toolset.

About

Hi, 👋, I’m Ryan Hefner  and I built this site for me, and you! The goal of this site was to provide an easy way for me to check the stats on my npm packages, both for prioritizing issues and updates, and to give me a little kick in the pants to keep up on stuff.

As I was building it, I realized that I was actually using the tool to build the tool, and figured I might as well put this out there and hopefully others will find it to be a fast and useful way to search and browse npm packages as I have.

If you’re interested in other things I’m working on, follow me on Twitter or check out the open source projects I’ve been publishing on GitHub.

I am also working on a Twitter bot for this site to tweet the most popular, newest, random packages from npm. Please follow that account now and it will start sending out packages soon–ish.

Open Software & Tools

This site wouldn’t be possible without the immense generosity and tireless efforts from the people who make contributions to the world and share their work via open source initiatives. Thank you 🙏

© 2026 – Pkg Stats / Ryan Hefner

nlp2terminal

v0.1.0

Published

Describe what you want in plain English, hit your hotkey, get a ready-to-run terminal command on your command line.

Readme

nlp2terminal

Press a hotkey in any terminal, type what you want in plain English, and the generated shell command lands directly on your command line — no copy/paste, just press Enter to run it (or edit/cancel it first, since it's a normal editable line).

How it works

nlp2terminal is one small binary with a prompt subcommand that:

  1. reads your one-line request from the real terminal,
  2. calls an OpenAI-compatible chat completions endpoint,
  3. prints only the generated command to stdout (all UI goes to stderr).

A tiny per-shell snippet (installed via nlp2terminal init <shell>) binds a key to a widget/function that runs nlp2terminal prompt, captures its stdout, and replaces the current input line with it — using each shell's own line-editing hook (zsh ZLE, bash bind -x, fish bind, PowerShell's PSReadLine). That's what makes "press Enter and it just runs" possible: the terminal never sees "output," it sees you typing a command.

Install

npm install -g nlp2terminal

(Or from source: npm install && npm run build && npm link.)

Configure the AI provider

If you already export an API key for a common provider — which most people do, for other CLIs/SDKs — nlp2terminal picks it up automatically, with a sensible default base URL and model, and you don't need to configure anything:

| Env var already in your shell | Provider | |---|---| | OPENAI_API_KEY | OpenAI | | ANTHROPIC_API_KEY | Anthropic | | GROQ_API_KEY | Groq | | GEMINI_API_KEY / GOOGLE_API_KEY | Google Gemini | | MISTRAL_API_KEY | Mistral | | DEEPSEEK_API_KEY | DeepSeek | | TOGETHER_API_KEY | Together AI | | OPENROUTER_API_KEY | OpenRouter | | XAI_API_KEY | xAI (Grok) | | FIREWORKS_API_KEY | Fireworks AI | | PERPLEXITY_API_KEY | Perplexity | | OLLAMA_HOST | local Ollama (no key needed) |

For anything else — a self-hosted vLLM server, a specific model/base URL override, or a provider not in the table above — set these three, either as real environment variables or in ~/.config/nlp2terminal/.env (see .env.example):

AI_PROVIDER_BASE_URL=https://api.openai.com/v1
AI_PROVIDER_MODEL=gpt-4o-mini
AI_PROVIDER_API_KEY=sk-...

AI_PROVIDER_* always wins per-field when set, so you can mix and match — e.g. keep using your existing OPENAI_API_KEY for auth but set just AI_PROVIDER_MODEL=gpt-4.1 to override the model.

Any OpenAI-compatible /chat/completions endpoint works.

Using a "thinking"/reasoning model? It writes out its reasoning before the answer — sometimes 1000+ tokens' worth — and if it gets cut off mid-thought you'll see an explanation instead of a command. Raise the response budget with AI_PROVIDER_MAX_TOKENS (default 4096):

AI_PROVIDER_MAX_TOKENS=8192

Verify it's wired up correctly — this also prints which provider/env vars were detected:

nlp2terminal doctor

Enable the hotkey

Add one line to your shell's profile, then restart your shell (or source it).

| Shell | Add to | Line | |---|---|---| | zsh | ~/.zshrc | eval "$(nlp2terminal init zsh)" | | bash | ~/.bashrc | eval "$(nlp2terminal init bash)" | | fish | ~/.config/fish/config.fish | nlp2terminal init fish \| source | | PowerShell | $PROFILE | Invoke-Expression (nlp2terminal init powershell \| Out-String) |

This installs two entry points in your shell:

  • Alt+K (Option+K on macOS) — live hotkey, edits your current line in place.
  • l2t <request> — type it as a normal command instead of using the hotkey, e.g. l2t list files over 5mb.

By default this binds Alt+K, not Ctrl+K — Ctrl+K is already kill-line in most shells/terminals and reliably collides with other tools, whereas Alt+K is conventionally unbound. Alt-combos are sent as the byte sequence ESC followed by the letter (\x1bk), which is how terminals represent "Meta" keys at the protocol level — not a Kitty-keyboard-protocol sequence, so it works on old and new terminals alike, on Linux and in Windows Terminal/PowerShell, with no configuration needed there.

Getting Alt+K (Option+K) working on macOS

macOS terminal apps generally treat Option+key as a way to type accented or special characters (e.g. Option+K types "˚" on a US layout) rather than sending ESC+k, unless the terminal is told otherwise. Pick your terminal:

iTerm2 — Preferences → Keys → Key Mappings → + → record Option+K → Action: "Send Hex Code" → value 0x1b 0x6b. (Scoped to just this one key combo, so other Option-based accented characters keep working.)

Terminal.app — Preferences → Profiles → Keyboard tab → check "Use Option as Meta Key". This is a blanket toggle (affects all Option combos, not just K), since Terminal.app has no per-key remapping.

Warp — Settings → Keyboard Shortcuts → add a custom binding for Option+K sending the escape sequence \x1bk.

kitty (~/.config/kitty/kitty.conf):

map opt+k send_text all \x1bk

Alacritty (~/.config/alacritty/alacritty.toml) — either set Option to act as Alt globally:

option_as_alt = "Both"

or scope it to just this key:

[[keyboard.bindings]]
key = "K"
mods = "Alt"
chars = "k"

WezTerm (~/.wezterm.lua):

config.send_composed_key_when_left_alt_is_pressed = false
config.keys = {
  { key = "k", mods = "ALT", action = wezterm.action.SendString("\x1bk") },
}

Usage

Live hotkey

Press Alt+K. A prompt appears right on your current line:

✦ find all png files over 5mb in this folder and list them by size

Press Enter once to generate — a moment later the line is replaced with:

find . -iname '*.png' -size +5M -exec du -h {} + | sort -rh

Press Enter again to run it, or edit it first, or Ctrl+C to cancel.

l2t as a normal command

$ l2t find all png files over 5mb in this folder and list them by size
  • zsh: the generated command is staged directly into your next prompt's edit buffer (zsh's native print -z) — just press Enter.
  • bash / fish / PowerShell: there's no equivalent built-in to pre-fill the next prompt in these shells (and the lower-level trick some tools use to simulate keystrokes, TIOCSTI, is disabled by default on most current Linux kernels for security). The command is staged into your shell history instead — press Up-Arrow, then Enter to run it.

License

Eligapris License (a modified MIT License) — free for personal and non-commercial use with attribution; redistribution, resale, and commercial use require permission from Eligapris. See LICENSE for the full terms.