It's true that a lot of adjacent tools in this space are built for non-developers. But if you try Retool, I think you'll notice that it's definitely not "no code" - any interesting application will require a significant amount of JavaScript to handle key interaction events and API/database integrations. The experience is closer to iterations of the Xcode Interface Builder where you'd visually design the UI/layout, then specify event handlers in Objective-C/Swift. It's much less code overall because you aren't doing boilerplate UI structure, styling, and low-level interaction event handling (focus and keyboard events, e.g.), but all the interesting stuff about your app still needs to happen in JavaScript.
Retool is still working out parts of the SDLC around modularity and the equivalent of a git flow for visual app development. Lots of energy going into this, would stay tuned for updates this year.
As you're grokking the mobile platform (or Retool generally) I think it's helpful to keep a couple things in mind.
- At least for now, Retool is designed and optimized for CRUD apps and operations software. On the mobile side, that will mean field workforce applications and mobile data entry stuff. If you need to provide a consumer-level user experience, building native (or React Native / Flutter at least) is probably your best bet.
- Consider solving problems with software on a spectrum from (approximately):
There is an appropriate time and place to deploy all these tactics. The sweet spot for Retool is for those use cases that can benefit from being codified in software (the flexibility of a more spreadsheet-y approach is detrimental/insufficient), but the business process is more important than granular control of the UI presentation. Instead of going full-on React, you can assemble a software-driven flow in Retool with about the effort it takes to assemble a non-trivial Keynote presentation (provided you know JavaScript - Retool is much harder if you are not a developer).
As always, there are trade-offs to consider when selecting the right tool for the job!
Look at the sparse amount of native components that comes with them. Need a native map view, or maybe AR? Overlay them with your own custom markers? Go native or at least React Native. One single tiny component that‘s missing will force you back to native code. And do you really want to invest months of work before realizing that you need to reimplement everything?
We are working on custom components for Retool Mobile. With custom components, we hope to give folks the ability to paint whatever component they need on the Retool canvas but there will of course be scenarios where retool isn't the right choice. We also build escape hatches where you can write JavaScript anywhere in Retool. We like to think we can cover a very broad base, especially in the scenarios in which Retool excels today.
Can you write well structured and reusuable code on retool? Can I write it in vscode, maybe even in typescript and deploy it up?
Most implementations of the "escape hatch" I've come across require scattering fragmented snippets across the project with no support for building out well structured and reusuable code.
I have to second this. Sandbox.io recently added VSCode integration (real-time sync with the code in the sandbox, which is now hosted in a micro VM), and this is a really good experience.
It used to be like that in the past, but now it's changing... For example Qonto, Spendesk - fintech unicorns, so rather serious companies - rely on Forest Admin. And if you talk to ops people using natively built internal tools, they typically complain about them a lot.
They are always missing features that most devs would refuse to live without...
Language features to support high reusability / modularity, robust dev tooling for navigating the software, support for proper a sdlc, etc...