It is really annoying if you have a package that is relatively new to the platform, and you type in the exact package name, that package is not reliably the first result.
Minor edge case, but infuriating if you want to check your own packages quickly (without needing to navigate menu > packages > YOUR_PACKAGE).
Still agree with you though, who is npmx actually for?
My only pain point, that I solved with a few lines of elisp, was to go to npmjs, to find the repo link for my dependencies. What's in `node_modules` may be compiled to an inscrutable blob and it's rare to find good library docs.
It sparks more joy than the old one and buddy, that's for frelling enough damn it. Whinge out!
Awful comment. Your comment is bad and should be ashamed.
Use it and disagree! Tell me it in fact does not spark more joy! This just seems pretty clear, ya'll.
It's just generally vastly nicer. I love that file exploration of packages doesn't feel like a last afterthought before leaving the solar system forever.
I don’t get it. What does it do? How can I trust it over simply visiting https://www.npmjs.com/ (which is perfectly fine to browse)?
Edit: it wants me to connect to the “atmosphere” - is this the Bluesky App Store thing? I really don’t see how linking my socials to npm search makes sense.
npmx maintainer here! a few answers to questions i see:
we haven’t launched yet! we’re building openly but aiming for march 3. as is par for the course, hn gets the scoop.
i made the first commit two weeks ago, so it’s very new. but we’ve had 900+ PRs and 170+ contributors in the last fortnight… because this is something we care about.
having said that, i’m taking notes about what is and isn’t intuitive
npmjs remains the source of truth for the registry, which is why we get our data from there
but along the way we add a lot of features, like:
- claiming new packages from the ui
- batch admin operations for your orgs, teams and packages
- total install size, vulnerabilities and deprecations for your transitive dependencies
- generated docs for packages
- linkable package contents
- and more
There’s also https://v1.run/ (packrun.dev) experimenting with an npm registry for agents that adds package scoring and MCP integration for AI coding assistants.
I've always defaulted to using https://yarnpkg.com/ to search for packages cause the npmjs.com search is so slow, but while the yarnpkg.com search is super fast, actually clicking on a package and seeing the details page takes forever.
This is super fast for both search and the details page, and it's super keyboard friendly which makes it even faster to use in practice. Definitely going to become my go-to search now. Love it, thanks for building it!
I'm not sure what features I'm supposed to notice that are better, but having built-in API docs and source code browsing is nice. (Though slightly laggy.)
Nit: there are distracting animations, such as on the weekly download graph.
The typeahead search speed is genuinely impressive. I was typing package names and the results were appearing before I'd finished the keystroke—that's the kind of responsiveness you usually only see with native applications.
As someone who spent a year obsessing over performance in a .NET MAUI app (IMAP processing, background execution, API latency), I know how hard it is to shave milliseconds off search. The npm registry is massive, and you're clearly doing something right on the indexing or caching side.
Are you using client-side filtering with a pre-built index? Server-side with some clever caching strategy? I'm genuinely curious about the architecture choices that got you to "uncannily fast" territory—especially compared to the official npmjs.com search.
Whatever you're doing under the hood, it shows. Performance is a feature.
I hate the trend of moving away from being able to search packages from the command line to searching inside a bloated web browser. It had a happened with PyPI. And now npm. Please stop this madness.
But why?
npmjs.com is not slow and not something I need to interact with very often.
And npmjs.com is still the authority when it comes to publishing packages, no? So I'd still have to use it.
Because they wanted to and Claude didn't tell them not to. Why even ask questions like this at this point.
In the age of LLMs I think we are going to see a Cambrian explosion of software.
Me personally, I’m writing tools for myself wouldn’t have bothered with before due the the time investment needed.
I almost never use npmjs.com.
When I do it’s just to click on the repo link.
It is really annoying if you have a package that is relatively new to the platform, and you type in the exact package name, that package is not reliably the first result.
Minor edge case, but infuriating if you want to check your own packages quickly (without needing to navigate menu > packages > YOUR_PACKAGE).
Still agree with you though, who is npmx actually for?
Whenever I know the name of the package I want to see, I always type in the URL directly: npm.im/[package-name]
npmjs search is very slow
how often does anyone use it though?
I mean are we really arguing over milliseconds here? I have never in my life had the thought "NPM search is too slow, I need a faster solution"
I have had the thought "NPM search sorted by downloads this week is giving me irrelevant packages" - but I'm not sure this tool solves that.
It can load for a couple of seconds (!) for me
My only pain point, that I solved with a few lines of elisp, was to go to npmjs, to find the repo link for my dependencies. What's in `node_modules` may be compiled to an inscrutable blob and it's rare to find good library docs.
It sparks more joy than the old one and buddy, that's for frelling enough damn it. Whinge out!
Awful comment. Your comment is bad and should be ashamed.
Use it and disagree! Tell me it in fact does not spark more joy! This just seems pretty clear, ya'll.
It's just generally vastly nicer. I love that file exploration of packages doesn't feel like a last afterthought before leaving the solar system forever.
Cynically, if you can attract a representative sample, you could aggregate and sell analytics data.
Another could be to have an "alternatives" section based on semantic similarity and / or some other features that have signal.
Here's one feature they have that seems important:
> It's good npmx.dev shows git and https dependencies. I still think it's crazy npmjs.org doesn't.
https://bsky.app/profile/dsherret.bsky.social/post/3mer2diwj...
I don’t get it. What does it do? How can I trust it over simply visiting https://www.npmjs.com/ (which is perfectly fine to browse)?
Edit: it wants me to connect to the “atmosphere” - is this the Bluesky App Store thing? I really don’t see how linking my socials to npm search makes sense.
The word “browser” in the title also tripped me to think about an actual browser
But it is a browser -- just not a web browser
directory perhaps less confusing.
npmx maintainer here! a few answers to questions i see:
we haven’t launched yet! we’re building openly but aiming for march 3. as is par for the course, hn gets the scoop.
i made the first commit two weeks ago, so it’s very new. but we’ve had 900+ PRs and 170+ contributors in the last fortnight… because this is something we care about.
having said that, i’m taking notes about what is and isn’t intuitive
npmjs remains the source of truth for the registry, which is why we get our data from there
but along the way we add a lot of features, like: - claiming new packages from the ui - batch admin operations for your orgs, teams and packages - total install size, vulnerabilities and deprecations for your transitive dependencies - generated docs for packages - linkable package contents - and more
Do you have your own dependency resolution algorithm? Can we have the most lightweight node_modules directory?
we do have an algorithm to calculate total install size
it’s not _exactly_ the same as if you install with npm and check the size of your node_modules but it’s a pretty good proxy
having said that, there’s room to make it better
There’s also https://v1.run/ (packrun.dev) experimenting with an npm registry for agents that adds package scoring and MCP integration for AI coding assistants.
Really fun / cool project!
I've always defaulted to using https://yarnpkg.com/ to search for packages cause the npmjs.com search is so slow, but while the yarnpkg.com search is super fast, actually clicking on a package and seeing the details page takes forever.
This is super fast for both search and the details page, and it's super keyboard friendly which makes it even faster to use in practice. Definitely going to become my go-to search now. Love it, thanks for building it!
Confusing name as npx exists (exexutes npm package without install)
- Certain pages load but are not able to load content, e.g. https://npmx.dev/package/@storybook/addon-docs fails to load content with:
> `[nuxt] Cannot load payload /package/@storybook/addon-docs/_payload.json?c459501f-8eb7-49c9-be9c-4a197fa35a39 Error: Invalid input`
- Scrolling fast on Firefox + Chrome is broken and resets the search results page to start.
- Pressing up/down arrows should navigate search item results instead of focusing individual tag elements.
I'm not sure what features I'm supposed to notice that are better, but having built-in API docs and source code browsing is nice. (Though slightly laggy.)
Nit: there are distracting animations, such as on the weekly download graph.
The typeahead search speed is genuinely impressive. I was typing package names and the results were appearing before I'd finished the keystroke—that's the kind of responsiveness you usually only see with native applications.
As someone who spent a year obsessing over performance in a .NET MAUI app (IMAP processing, background execution, API latency), I know how hard it is to shave milliseconds off search. The npm registry is massive, and you're clearly doing something right on the indexing or caching side.
Are you using client-side filtering with a pre-built index? Server-side with some clever caching strategy? I'm genuinely curious about the architecture choices that got you to "uncannily fast" territory—especially compared to the official npmjs.com search.
Whatever you're doing under the hood, it shows. Performance is a feature.
Whats even crazier is if you click on thr author link, there other packages are listed almost instantly. Crazy damn speed. Beautiful to look at too.
Don't use an LLM to write comments for you.
If I scroll too far down the search results it snaps back to the top. The typeahead search is almost uncannily fast though.
I hate the trend of moving away from being able to search packages from the command line to searching inside a bloated web browser. It had a happened with PyPI. And now npm. Please stop this madness.
What has moved away? You can still use `npm search` just fine.
cool
I like it, don't get the negativity in this thread
I think it looks cool. I like the mouseover effects. I still echo others concerns e.g. npmjs.com is the ultimate authority of what NPM packages exist.
Did you create this? If not how did you come across it?
Aaaand, it's down. Is it's inability to handle light load what makes it modern?
it’s modern, i heard