Insomnia 8 forces users to login and use cloud storage

Someone just left a comment suggesting Restfox, so that's probably where I'm jumping to next:

Time to fork.

If someone wants to create a product that doesn't store your data externally I know many people willing to pay for it.

Not affiliated with any of these projects, just anecdote. My journey has been Postman => Insomnia => Restfox. Just made that third jump today and everything ported over fine and looks to be working great. 🤷

https://github.com/Kong/insomnia/issues/6577#issuecomment-1738052455

Some people are discussing alternatives in a comment thread on the hackernews too:

• Hurl, a command line tool that runs HTTP requests defined in a simple plain text format.
• Bruno, a lightweight Postman alternative
• ThunderClient, a vscode extension

I don't know your usage of REST clients, or more precisely the other tools you work with aside this tool, but in the Jetbrains IDE suite, there is a HTTP client included which is used by writing requests inside a .http file, like this:

GET http://127.0.0.1/get-something

###

POST http://127.0.0.1/create-something
Content-Type application/json

{
  "some": "data"
}

It has a miriad of benefits right out the box:

• The fact its simply text files means you can track them via git, it's impossible to "lose" your requests via a borked update of your tool, it will remains as a file on your filesystem that you can read easily.
• The way they are written is practically 1:1 the same as what your TCP request will send down the wire to your server.
  No magic of the tool trying to understand what you want to send when you send files (Do you want to send them via a Form-Data or directly its content. Do you want to customize each parameter in your Form-Data, etc...).

# This is clearly a form-data request
# And you can also see how a form-data is constructed in raw form
POST http://127.0.0.1/post-file-with-some-metadata
Content-Type: multipart/form-data; boundary=Boundary

--Boundary
Content-Disposition: form-data; name="metadata"
Content-Type: application/json

{ "user_id": "646161", "ts": 156113130 }
--Boundary
Content-Disposition: form-data; name="profile"; filename="98461613146161.jpg"
Content-Type: image/jpg

< ./the_image_is_algonside_this_file_and_I_can_add_it_via_its_path.jpg
--Boundary--

###
# While this is sending the file directly
POST http://127.0.0.1/post-file
Content-Type: image/jpg

< ./the_image_is_algonside_this_file_and_I_can_add_it_via_its_path.jpg

• The IDE still have many niceness, like working with cookies is done automatically by default but you can deactivate them with a comment # @no-cookie-jar just for the request under it.
  You can setup variables with different values depending on what they call "Environment" that you can switch in the GUI.
  In practice I use it to have all my requests written like this:

GET {{server1}}/get-something

And in my rest-client.env.json file at the top of the project I have:

{
  "local": {
    "server1": "http://127.0.0.1:7000",
    "server2": "http://127.0.0.1:8080"
  },
  "production": {
    "server1": "https://example.com",
    "server2": "https://otherdomain.com"
  }
}

So I can switch between my servers and the production ones with one switch in the GUI, while keeping all written requests working.

But the main stopping point, is do you already pay for the JetBrains suite (I don't think it's available in the IntelliJ Community Edition, but I could be wrong) ?

If not, I think there is a plugin for VS Code which seems to work with .http files, but I've never used it.

Ugh I'm in the exact same boat, and did the same about a month ago. I had almost used Milkman since it looked like a clone of Postman, but I figured I would be better off with something with a bit of corporate backing instead of a single developer's pet project. Turns out that that was a mistake. I see recommendations in comments for Restfox as well, which seems to have a few more github stars and such on Github.

Someone's gonna give the asshole suggestion, so it might as well be me: use curl. ;)

(I do this. I use shell history to collect requests. The downsides are pretty obvious, lol, but there are some upsides: it's very clear exactly what request you're making; you can customize the request in extreme detail; documentation is good, and online guidance is extremely extensive; and, to my eternal amusement, "curl commandline" has become a standard request description format. And, I guess, the risk of enshittification is very, very low. I dunno, I do it because (a) it's shell-scriptable, and (b) I really hate Postman.)

I've been trying out the REST extension for VS Code, which lets you define behaviors in .rest files. The code sits in the repo with the API it's testing, which is nice. The overall experience is not super smooth, and I don't love having yet another tab crowding my code window.

Honestly, I mostly just use curl unless there's an annoying sequence like fetching a cookie to attach with every request.

We evaluated their Kubernetes gateway product a few years ago. We laughed them out of the (virtual) room once they told us pricing and we just went with Istio. The whole experience with them left a bad taste in my mouth, where I almost didn't use their API client after dumping Postman. I should've gone with my gut.

This seems to essentially be the FAQ subnetmarco, Kong CTO, posted in the thread yesterday: https://github.com/Kong/insomnia/issues/6577#issuecomment-1738131554