hckrnws
Show HN: Crust – A CLI framework for TypeScript and Bun
by jellyotsiro
We've been building Crust (https://crustjs.com/), a TypeScript-first, Bun-native CLI framework with zero dependencies. It's been powering our core product internally for a while, and we're now open-sourcing it.
The problem we kept running into: existing CLI frameworks in the JS ecosystem are either minimal arg parsers where you wire everything yourself, or heavyweight frameworks with large dependency trees and Node-era assumptions. We wanted something in between.
What Crust does differently:
- Full type inference from definitions — args and flags are inferred automatically. No manual type annotations, no generics to wrangle. You define a flag as type: "string" and it flows through to your handler.
- Compile-time validation — catches flag alias collisions and variadic arg mistakes before your code runs, not at runtime.
- Zero runtime dependencies — @crustjs/core is ~3.6kB gzipped (21kB install). For comparison: yargs is 509kB, oclif is 411kB.
- Composable modules — core, plugins, prompts, styling, validation, and build tooling are all separate packages. Install only what you need.
- Plugin system — middleware-based with lifecycle hooks (preRun/postRun). Official plugins for help, version, and shell autocompletion.
- Built for Bun — no Node compatibility layers, no legacy baggage.
Quick example:
import { Crust } from "@crustjs/core";
import { helpPlugin, versionPlugin } from "@crustjs/plugins";
const main = new Crust("greet")
.args([{ name: "name", type: "string", default: "world" }])
.flags({ shout: { type: "boolean", short: "s" } })
.use(helpPlugin())
.use(versionPlugin("1.0.0"))
.run(({ args, flags }) => {
const msg = `Hello, ${args.name}!`;
console.log(flags.shout ? msg.toUpperCase() : msg);
});
await main.execute();
bun create crust my-cli
Happy to answer any questions about the design decisions or internals.
This looks useful. But, it's interesting how the backend-world and front-end world keep diverging. I must admit, I had no idea what this was from the title. "CLI framework"? But in backend-land, these would typically be called "argument parsers" or "command line argument parsers". But maybe I am missing some of the functionality.
Both in the frontend and the backend, I've usually used "If it calls your code, it's a framework, if you call its code, it's a library", and would seem to fit here too. An argument parser you'd call from your main method, then do stuff with what it returns. In Crust, it seems you instead setup the command + what will happen when it's called, then let the framework call your code.
That’s… actually a great definition. I’m going to try to retain that.
This is classically called "Inversion of Control"[0] or the "Hollywood Principle"[1] as in "Don't call us-- we'll call you".
[0] - https://martinfowler.com/bliki/InversionOfControl.html [1] - https://wiki.c2.com/?HollywoodPrinciple
good point.
we’re using “framework” intentionally because it goes beyond argument parsing. crust handles parsing, but also:
type inference across args + flags end to end compile-time validation (so mistakes fail before runtime) plugin system with lifecycle hooks (help, version, autocomplete, etc.) composable modules (prompts, styling, validation, build tooling) auto-generates agent skills and modules from the CLI definitions
so it sits a layer above a traditional arg parser like yargs or commander, closer to something like oclif, but much lighter and bun-native.
Hi, I called it "CLI framework" because it is more of a ecosystem of modules that contains everything you need to build a CLI, argument parsing is just part of it. The @crustjs/core module is the argument parser, and there are more modules such as @crustjs/skills that would derive agent skills from your command definition, @crustjs/store that state persistent and so on
I was looking for something like this.
Does the help plugin not support color?
Looks like the spacing is messed up too. I just converted my app to use it and its coming out like
``` COMMANDS: export-schemaExport table definitions from existing database to YAML export-dataExport table data in CSV format import-dataImport table data from CSV file schema-sqlConvert YAML schema back to MySQL export-usersExport users in YAML format users-sql Convert users.yaml back into SQL export-allExport all data from host export-all-tgzExport all data from host databases-sqlConvert databases.yaml back into SQL export-typescriptExport TypeScript interfaces ```
i.e. there's no space after `export-schema` it just goes immediately to the description.
I will add color support for help plugin later. open a issue if you found bugs or improvement
also: no "C", and no "rust", despite being a portmanteau of both these other languages names :)
Looks cool. If anyone is interested in a simple option with autocomplete working out of the box and no extra bells and whistles, feel free to check out my project: https://github.com/gutenye/script.js
The plugin system and compile-time validation are nice touches. My main concern would be the standalone binary size — 58-109MB is a lot for a CLI tool. Most Go CLIs land around 10-20MB and even that gets complaints. Is there a way to strip unused features, or is this mostly the Bun runtime being bundled?
The binary size is my main hesitation with Bun-based CLIs too. For comparison, a typical Go CLI with cobra compiles to ~10MB statically linked, and you get cross-compilation for free. The plugin system and compile-time validation here are genuinely nice though — Go's CLI ecosystem doesn't have anything equivalent to that.
this is cool! i'd recommend fleshing out the README. Clicked on the link before the discussion and was a tad confused.
Hi, I will update the README.md to make it more informative. The reason I left it kinda empty is because curst has a website that have more details about this project at crustjs.com. I just had the link in README.md for now
will fix in the next hour!
Any plans to support Node.js? Also some comparison (at least design choices) with existing frameworks would be nice.
Technically most modules would work with node.js as they are not using any bun specific APIs. The reason for crust to be all in bun is because bun can compile your cli into binaries for distribution which powers curst build. The idea is that since you will have bun bundled with your CLI to end user. the developer does have to worry about the end users don't have bun installed and use Bun api freely when building.
Is there an examples section? Would be helpful to see a demo
one of the examples would be trynia.ai (search and index api for ai agents)
here is github: github.com/nozomio-labs/nia-cli
nice, congrats on launch. To get an idea... what's the size of a standalone hello world cli binary?
tens of KBs (v small)
Isn't a standalone Bun binary like 50MB because it has to bundle the runtime? How could this get smaller?
Hi, the creator of crust here, the binary size varies between platforms. With the hello world cli, the smallest, on darwin-arm64 it is 58.1M, the largest on windows-x64 is 109M. hope this helps!
seems pretty useful and easy to setup
I’ve been using the jsr:@cliffy/* packages from deno to solve the same problem.
yup yup same. That one's worked well for me. Between that and the deno std, it's nice to have it feel like mostly everything you need is available with very little searching.
> Versions before 1.0 do not strictly follow semantic versioning.
Sorry for being nitpicky, but yes they do. Semantic versioning[0] allows arbitrary changes while the major version is 0:
> Major version zero (0.y.z) is for initial development. Anything MAY change at any time. The public API SHOULD NOT be considered stable.
[0]: https://semver.org/
Hi, the comment I left in README.md is just to warn user that do expect breaking changes before 1.0 for THIS PROJECT specifically. there is no implication that semver as a standard is promoting arbitrary changes/version before 1.0.
I get that, and it's a nice thing to warn users about. My nitpick was merely about the specific "do not strictly follow semantic versioning" phrasing. There's nothing you can do in a 0.y.z release which doesn't follow the SemVer spec (e.g. your next release could turn the project into a spell checker instead of a CLI framework and you'd still be in compliance).
ohh thank you for pointing that out. I just double checked the semver docs and it makes sense now. Something new I learned today
thanks for the catch, what we meant is that we’re not committing to strict stability guarantees yet, so APIs may still change as we iterate toward 1.0.
I understand, but that's already implied by a 0.y.z version number.
I will say that this doesn't seem to be how semver is used in the wild, which I would argue is more important. I personally didn't know about this rule. Tons of Rust projects follow semver don't follow it either, and just stay on 0.x.y forever.
How is it used in the wild, in your experience? Have you observed projects following some alternate set of rules?
I thought projects that stay on 0.x.y forever mostly do it because it means they're allowed to break things. Also, since 0.x.y means "anything goes", projects can introduce their own conventions within that range without violating the spec.
I know that some package managers (including Cargo and npm) confusingly treat 0.1.0 → 0.1.1 like a "minor" update, despite the spec. Is this what you're referring to?
wow, cannot wait to start with this
[dead]
[dead]
Psst, the GitHub link in your post is broken (it should be https://github.com/chenxin-yan/crust).
Fixed above. Thanks for the heads-up!
thanks for flagging! the post itself works, just the link at the bottom
Comment was deleted :(
Crafted by Rajat
Source Code