Why API Tooling Still Feels 10 Years Behind
Posted by Successful_Bowl2564@reddit | programming | View on Reddit | 14 comments
Posted by Successful_Bowl2564@reddit | programming | View on Reddit | 14 comments
programming-ModTeam@reddit
This post was removed for being off topic for r/programming.
Farhaan_1120@reddit
Honestly, it feels like each API tool solves one problem and then creates two more. You either get great documentation but poor DX, or nice SDKs that break every few versions.
The weird part is frontend tooling evolved so fast in the last 10 years, but APIs are still stuck with half-baked standards, inconsistent error handling, and “just read the docs” as the main solution.
At this point I care less about features and more about stability and predictability. A boring, well-documented API > a “modern” one that changes every 6 months.
GuaranteePotential90@reddit
hah indeed, I dont wanna hear more "modern".
shahara_10@reddit
Same 😅 ‘modern’ these days just means breaking changes with better marketing.
GuaranteePotential90@reddit
Well yes. When I was growing up, modern meant flashy with strass and blinking lights. Shiny.
At some point, still today, modern means "using AI".
In this context, for dev tools, I think modern means staying out the freaking way and be something developers use "in the flow of coding".
dayk995@reddit
There’s a lifecycle to these tools. They prove they’re useful, get bloated with niche features, acquired and then become unusable. Rinse and repeat.
With AI it’s actually super easy to create these tools for your own needs. I created a simple webocket client where I can test sending/receiving messages in just an hour or two. I can share the repo if anyone is interested.
GuaranteePotential90@reddit
did you open source the websocket client? Is it 100% vibe coded?
dayk995@reddit
Yeah, I’ll admit it vibe coded. It’s nothing terribly impressive but I needed to mock and test things quickly and easily without having to run the actual client (a game in my case).
https://github.com/dimxasnewfrozen/patchbay
Foreign_Yogurt_1711@reddit
Because the people building the tools aren't the ones suffering through them daily. Postman had one good era then got acquired-brained into an enterprise collaboration platform nobody asked for. Insomnia got quietly killed by Kong. Bruno is promising but still rough. So everyone just stays on the bloated thing they know because switching costs are real and the new thing might get ruined by an acquisition in 6 months anyway. The deeper problem is that this kind of pain is invisible. Bad tooling doesn't cause outages, doesn't miss deadlines, doesn't show up in a postmortem. It just silently steals 30 minutes from every developer every single day and never makes it into anyone's roadmap because you can't point to it bleeding. And the documentation situation is its own separate tragedy. The gap between what the Swagger spec says and what the endpoint actually does in production is where entire afternoons go to die.
yojimbo_beta@reddit
Sometimes I wonder about writing my own tooling, and sharing it. I would be happy to maintain a no-bullshit, no-monetisation API client. But, how would I even promote it? There are so many postman clones
GuaranteePotential90@reddit
why dont you join this one? Built this with a few folks: https://github.com/VoidenHQ/voiden and open sourced it recently so that the community owns it. We have been working on this and built it for us internally initially (in my company we build 100s of APIs every month) but then we kept refining it and it slowly became something that 1. we loved and 2. we have no interest in monetizing since we all have different jobs and focus. So this is intended as a tool that will stay free, light and will bring real devs that do real API work. (apologies for the small "pitchy" vibe but It would be great to get more folks into improving this space).
yojimbo_beta@reddit
Oh, interesting. I will look into that!
Foreign_Yogurt_1711@reddit
The promotion problem answers itself honestly. Build it, put it on GitHub, write one honest post about why you built it and what you refused to add. Developers are so burnt out on bloated tools that 'no monetisation, no bullshit' is genuinely a differentiator now. The Postman clone graveyard is full of things that tried to compete on features. Nobody has really won on just staying minimal and trustworthy.
macrohard_certified@reddit
Reusing pieces -> use variables or scripts to capture values.
Tracking changes -> API responsibility, can be done with proper docs, Git and API versioning.
Writing docs -> API responsibility, not consumer's.
Running tests -> integration tests shouldn't run on API testing tools, read this.