Without sounding like every other developer who hates on electron, I would really appreciate an http client with a Gui that was lighter, like iced or slint.
I appreciate the versatility of electron and it giving us beautiful and usable apps but I have 16 GB of ram, I can't upgrade it and I genuinely have multi second hangs after I have 3 instances of VSCode, Firefox and Chrome open along with Bruno.
> Without sounding like every other developer who hates on electron, I would really appreciate an http client with a Gui that was lighter, like iced or slint.
How much would you be willing to pay for such a client?
I'm not being facetious, I'd really rather like to know, because as an independent developer, every single time I suggest a native application using native widgets to a client, they choose the HTML interface rather than a Qt or similar based interface.
The easily doable pretty animations in CSS in a 800MB-in-RAM-while-running application is, to paying clients, preferable than a 50MB-in-RAM-while-running that doesn't have the fancy spinning, tilting, animated wizz-bangs.
I have been writing software professionally since the mid-90s; I can do you a quick GUI-based cross-platform HTTP send+receive application based on libcurl in about 2 weeks. Looking at the minimum I need to make to pay my bills, I need purchasers paying a cumulative 1000USD for this effort, so 10x buyers @ $100, or 100 buyers @ $10, and so forth.
And, of course, I'd expect to only be able to sell it for a short while, if it is popular, until a clone starts up with $40k worth of SEO and advertising money.
The software you want can be had, and the skill to make it exists, in a timeframe that is feasible, but the economics are just not there.
This is a good business perspective, but I don’t really see why this couldn’t be someone’s open source passion project. The comment wasn’t really implying that they needed a paid tool, just a tool that suits their needs and is lightweight. Plenty of the software I use on a daily basis is open source software where you could argue that the economics shouldn’t have been there.
> Plenty of the software I use on a daily basis is open source software where you could argue that the economics shouldn’t have been there.
Me too. But the problem with "the economics just aren't there" means that if I cannot get, just from word-of-mouth (say, a Show HN post) 100 users @ $10 once-off lifetime purchase, then this is not a product that is in demand anyway. An open-source/free product that is exactly the same would similarly receive no love from users.
IOW, if not enough users exist for this product at $10, not enough users exist for this product at $0. Your passion product will still result in the dev burning out on the fact that no one wants their passion enough.
Doesn’t that depend on the buyer? I can think of several products that I would only use if free, and would go without if it was $1, $5, or $10. E.g. todo list apps, time tracker apps, budgeting apps.
I think you can argue that, if you have enough demand at $10, that you’ll have enough at $0, but I don’t see how not having demand at $10 implies that you won’t have demand at $0, since usually making something cheaper can change buyer’s minds.
> Doesn’t that depend on the buyer? I can think of several products that I would only use if free, and would go without if it was $1, $5, or $10. E.g. todo list apps, time tracker apps, budgeting apps.
My point is not that the user would go without. My point is that if your product is not desirable or differentiated enough that enough people would pay $10 for it, then making it free won't help because it will be lost in a sea of clones that already existed before you even started working on your product.
After all, look at your list - todo list apps, time tracker apps and budgeting apps; you cannot, in the sea of free competition right now, deliver a desirable enough app in any of those classes, unless it provides so much value over and above the others that many people are willing to shell out $10 for it.
IOW, if the product does not provide enough value for some people to pay $10 for it, making it $0 won't make a difference because it doesn't provide anything over and above the entrenched free offerings.
I find this funny from people that make their living from $$ from software! Lol
But apart from that Bruno is open source (MIT), you can fork it and have it anyway you want…
The paid option seems just like the traditional open source model of make software free but offer enterprise support. And given that CTOs tent to want to pay for stuff (postman is terrible nowadays but is still picked and paid for) why would Bruno straight up say no to money?
Open-source is about having the whole stuff, not restrictions on "Embed Bruno into your SDLC with deeper Git integration, automation."
What I do resent is the fact that Bruno was a reaction to shitty restrictions from another application, and they did the same thing a few months/years later.
I'm not sure how adding a subscription is 'the same thing'. Postman forced all of your data to be stored in their cloud, and held it hostage for $1,000/user/year.
Bruno seems to charge somewhere between 70-130 dollars user/year for unlocking GUI features. The same features are all accessible via the terminal or IDE.
https://github.com/usebruno/bruno has the source code. It's a node app so i guess by binary you mean a way to run it without the normal electron wrapper? You should be able to run it standalone though i've never tried it outside it's normal distribution method.
Why is webview a problem? I hear this a lot but not sure why (with the exception of gtk WebKit on Linux which has legit perf issues). We’re on web right now and I’ve never heard anyone complain hackernews is sluggish and that they want a native app instead (or rather 5 native apps minimum for the big OSs).
This is mostly the case because most solutions provide more than the bare minimum of DOM rendering and event binding that a web view originally entails. Once you "accidentally" ship an entire browser inside your app, you've opened up more vectors for vulnerabilities—such is the price of humanity's hubris in attempting omnipotence.
Then second aspect is the "well-hidden" JS runtime or the general dislike of Javascript, but this point has been explained by other commenters well enough.
> Once you "accidentally" ship an entire browser inside your app
That’s not needed. Generally there’s a webview available on the system of choice. All major platforms have it, including mobile and many Linux distros.
> vulnerabilities
Such as? I mean yes if you load remote content with local access to FS etc (although that’s not within the webview). But you don’t need to (nor should you).
> Without sounding like every other developer who hates on electron
Electron is hated for very good reasons. Postman in particular is just so insanely bloated and sluggish, it's painful to use on anything that doesn't have a higher end CPU.
My whole organization picked up Bruno after the whole Postman fiasco. I've found Bruno to be very unpolished and buggy. We've run into bugs where the UI will show one value, but when the request sends it uses a previously overwritten variable.
Looking forward to having to move to something else, again.. sigh
It doesn't seem like the worst - yet - but Bruno also ramped up their monetization. So far it seems survivable. But given how things have gone before, it's unclear how viable Bruno will really remain.
Yeah I started using the GUI version, there’s a standalone app or the in browser one. I don’t know why all of these web request clients have been trying to sell enterprise features instead of just giving you a gui to make a quick request.
Hmm, I didn’t hear about this! What about that lifetime Bruno thing that was $9? I’m pretty sure I bought that. Is that just nothing now or what does that get me?
Bruno's pricing has changed recently, the lifetime offer is not available anymore. Prices tiers are: free, $6/month, $11/month [1]. Bruno's developers explained it was necessary to sustain Bruno's growth.
It's an http client... Itonically I think curl has grown to become so pervasive that it's common parlance to "curl something" to a large extent because haxx thankfully has a radically different approach.
this is (at least) the third time we've seen this in this exact space: Postman, Insomnia, Bruno.
At this point I just use the REST client extension in VS Code (Rider has one too) or HURL. The lack of GUI makes it a little tougher for new people, but file-based is nice, and in the end I have much stronger skills & understanding in the area.
> Bruno is offline-only. There are no plans to add cloud-sync to Bruno, ever. We value your data privacy and believe it should stay on your device. Read our long-term vision here <https://github.com/usebruno/bruno/discussions/269>.
I glanced over that github discussion and don't see where they've gone back on that statement. Am I just missing where they've taken the cloud route?
I’ve been burned a few times by these clients — difficult backups, changing licensing/commercial terms, hard to version — so now I prefer a few simple .http files that I can version in Git and easily read, even if the extension disappears.
I've landed on the REST extension in VS Code for development and interactive work, and Hurl for more test/automation stuff. Both take a little more effort to get setup but way more productive than the typical Postman workflows I see.
Idk if they have it yet, but last time I used the VS Code REST extension it still didn't support pre/post request scripting (which is already insanely fragmented syntactically across Postman/Insomnia/Bruno).
For any long-term things that would normally be in a Postman collection and distributed to others / outside teams, I usually end up with folder-separated Hurl scripts (similar to your use).
I can totally see how the REST extension is way easier for point & shoot from within VSCode though!
If you're just _hacking_ a few simple calls, curl is the way to go.
But if you're working in a team, with multiple environments, with complex payloads, authentication, doing dozens of API calls everyday...
Having a software able to manage libraries of endpoints, parameters, simple environment switching, included auth, sharing between team members... is a big time saver.
I personally prefer IntelliJ's HTTP Client[0] since I always have my IDE open, the files are not obfuscated in a gibberish format and can be versioned/shared through Git.
But when I start working on an existing project, having a Postman collection to rely on is a huge time-saver, instead of having to go down in-existent API docs or trying to infer from the code itself.
This and also when you newly join a team it is more productive to start using the tooling what they are using and move to preferred tooling once familiar with the API endpoints.
You may like Hurl [1]: it's an Open source cli based on curl (libcurl to be exact), to run and test HTTP requests with plain text (I'm one of the maintainers). We recently add --curl option to come back to curl. Give it a shout!
I've recently used Hurl to create a test suite for migrating a complex Nginx configuration to Caddy and it was a great choice!
I ran Caddy replacing the upstreams with mockbin-like services (don't remember which one I used) so it would respond with information about the request made by the proxy, so Hurl could make assertions on that.
Recently tried out hurl for a project to show how abstract tests can be run in a specific environment. Great tool, it will definitely stay as part of my standard toolset.
I cannot express in words how much as an engineer I am sick of every app people suggest to me needing an entire quasi-visualized OS running behind it, written in the shittiest language ever to grace our cursed machines, just to render text and perform web requests.
This is one application where I simply prefer a GUI for most of the use cases. All the various components of an HTTP call are visually better represented as far as I am concerned. Often enough it is also specifically the payload on both sides I am specifically interested in and viewing them side by side and making quick adjustments is just my personal preferences.
In the year 2024 with 8 cores, 32gb of ram and 4tb of storage available the electron overhead generally also doesn't really matter that much to me.
Though this doesn't seem to be an electron app either, rather a PWA. Which makes me wonder how well it works with all CORS limitations you are facing in the browser.
Not that I have anything against CURL either, when I am working in a CLI (for example on a server) it is the perfect tool.
I use a jupyter notebook and Python with requests.
The problem with all these tools is you pretty quickly end up basically wanting full programmability...at which point, Python is just easier and more flexible.
Combine that with uv for on-demand script dependencies and it's also completely shareable and manageable.
Curl is nice but its syntax leaves a lot to be desired. It's also hard to distribute to others a "collection" of them, as you're going to end up with troubleshooting questions from less-technical users, especially when it comes to multi-request items that require pre/post response processing.
No, you are not the only professional among the amateurs/hobbyists.
I know that sounds offensive, but let's be real, nothing wrong with using AI to code or using a gui to learn about the underlying protocol. But it is what it is
I always use curl, jq, etc. It's so simple and straightforward. But I would be lying if I claimed it to be easy to convince my teammates to use the same toolset I use.
I've run the open source version of Insomnia (https://github.com/ArchGPT/insomnium) for a while, but as that project is no longer actively maintained, I may give this a look.
He sold it some time ago. Either it expired, he negotiated out of it, or there was never much of one to begin with. My guess it was a little of all three. I think a lot of what Kong wanted with it already happened - it seems to have helped them grow. They still have it and it's still helping but I don't think Yaak is that big of a threat to it. Kong still has a major API client in a space that has room for more than one.
It feels like they were after a similar strategy as when Fiddler was sold. They can control the narative and likely use it as an onramp to their other (much more expensive) products when developers introduce any of their tolling into the enterprise.
I believe both Bruno and Hoppscotch are "open core" with expensive paid plans for additional/enterprise functionality, and they limit community-contributed functionality in the open source versions to avoid competing with the commercial versions.
The app is called “RapidAPI” after it got bought, now apparently the team got bought by Nokia so would not buy into it too much at this point. It used to be a nice native app.
What is https://httpyac.github.io/ missing in comparison to postman, insomnia or hopscotch? I really like being able to run .http files as unit tests in ci as well as during debugging.
A big plus imo is that you can check in and version control the files alongside your application. Very convenient when working on micro services in a team.
Going to leave this here for shady practices. A pull request was declined by the CEO since they were planning to build an OIDC feature into the enterprise plan.
This is really weak. I'm all for benevolent dictators ruling their projects to protect the conceptual integrity - until that means anti-user behaviour to only enrich themselves.
"But we need to do this to survive!"
no you don't; every handy tool with many substitutes doesn't need to be a startup, especially when your competitive difference is temporarily not doing the thing you're now going to do.
That's incorrect. JetClient has a free tier that doesn't require a subscription, an email, or a credit card. You can even install it on the free IntelliJ Community Edition without a JetBrains account. Only the Pro tier requires a subscription—just like most API clients. But the free version includes many features and is comparable to Bruno's free tier, not just basic functionality.
One thing I really like in the SoapUI, is the ability to have mocks(for both REST and XML), so I can mock the responses of some of the requests. Any of those other opensource alternatives have that option?
so I get I can run it offline, and call local (dev) apis, but is this really the major feature we want & need? How are people going to react when he pulls the same bait and switch for their API tooling a second time?
I appreciate the versatility of electron and it giving us beautiful and usable apps but I have 16 GB of ram, I can't upgrade it and I genuinely have multi second hangs after I have 3 instances of VSCode, Firefox and Chrome open along with Bruno.