Monthly Community Update - January 2026

[kev.nobre]

I’m actually getting a little discouraged already… I don’t think you guys understand that no one who already knows Bubble uses this AI Agent, and I’ll leave some of the reasons below:

1- Those who work with Bubble use it precisely because they want to do something customized. They already have the project, the designer, the database, the structured logic, and the direction in which to scale the app. If I try to create this with AI, it won’t meet these demands 100%, because there are many minute details, and with the implementation of WUs, application optimization is essential, as are some tricks that Bubble has up its sleeve for this, which AI won’t apply itself.

2- The agent only works for apps initially created with AI, so it won’t make much difference for existing apps. Of course, every day someone is making a new app, but those who already work with this are already used to doing it without the AI Agent. and even if they do, as I said above, most get lost after the user fixes some bugs and tries to change the designer. Unfortunately, AI is not viable for Bubble. I understand that it’s because of the market, but it’s not something scalable for your business idea!

Moving on to a new topic, I’m going to talk about new updates for developers. A little over a year ago, you announced the implementation of native branches (if/else) loops, and other ideas such as a goTo, for example. Is it really that difficult to add these basic utilities that we use every day in programming? It’s depressing to get into the workflow and have to use an API to do something that should be so simple, or simply duplicate a button to be able to do two different conditions because of a simple if/else.

Creation of new native group models, PDF creation, EDITING THE DROPDOWN MENU DESIGNER, among others. I know there are plugins, but you know that if one day the plugin crashes, we’re the ones who get screwed, because they’re not native, and that’s sad for the user’s day-to-day experience. And what about editing the designer on the input menus? Come on, it should be mandatory, because each menu is ugly and can’t even be edited with CSS/html.

Improve the dynamics in the URL parameters, because it is not possible to leave this dynamic in the Workflow, having to write it by hand every time, and if the app is very large, changing this by hand will drive the user crazy.

So that’s my current experience and my honest opinion. Finish this AI Agent soon and prioritize those who are already bringing money to your platform, not those who may only be potential future customers!

[senecadatabase]

I think once Bubble gets its AI where they want it to be, it will be the top in its class.

It’s not there yet, but I can see it’s going in the right direction and will be there soon.

Congratulations to everyone at Bubble for doing such a good job.

You’re growing the right way, and it will get there

[sandhya.jaideep]

Completely agree. The ability to branch in workflows would be awesome. I’m tired of hacking it out with separate workflows with conditions attached.

The lack of a better debugger with the ability to debug backend workflows is my pet peeve. The current debugger is just not enough.

[boston85719]

@josh

API workflow permissions: We’re adding an option for backend API workflows to be runnable only by users with admin API keys. Users will be able to select if API workflows can run without authentication, only by authenticated users, or only by admins. This update should be live in the next week or so.

Does this update solve for this idea? Run As User Action for Backend Workflows

[code-escapee]

@davidthorpie

The three features that would be truly game changing:

• Ability to choose async/sync workflows (and ideally functions) - this is by far the most infuriating aspect to bubble is trying to work out under what conditions things run, as it is dynamic. At the very least, TELL me how they will run to save me the hassle!
• if/else conditional branching
• Loops

+1. Been asking about loops and real if/else branching for well over a year now. At this point it honestly feels like we’re being gaslit because every thread gets radio silence. (Unless I somehow missed an update?) Can we just get a straight answer on whether this is still planned or if it quietly died? Even a simple one-line status check-in would help.

[code-escapee]

Thanks Cory, appreciate the clarification.

You’re right that I was blending the Agent with App Gen, that’s fair. But even focusing just on the in editor Agent for AI apps, the core issue feels the same.

The question isn’t which surfaces the Agent can touch, it’s whether what it produces actually survives real building. Most generated UI still gets rewritten once you’re dealing with real data, edge cases, and iteration.

Given its limited utility, the “full-stack” framing feels premature.

If the Agent got genuinely good at that layer first, it would change how people actually build.

[mike.alvarez]

I can say that workflow improvements, loops and branching are all my radar and being considered for H1 initiatives.

[davidthorpie]

This is great news if it comes true! List Shifter plugin has been for a LONG time the only solution for loops. I hope this conditional logic doesn’t consume WU though, and I think bubble really needs to consider a more graphical workflow editor to make it very clear how the branches work. It will also need to be super clear what is running client vs server side, as I cash see this logic causing lots of confusion and excessive WU consumption if not well understood. Either way, glad it’s on the radar finally!

[randomanon]

@mike.alvarez

loops and branching are all my radar and being considered for H1 initiatives.

Really need this. Hopefully parallelization too, which is even more important than those things!

[code-escapee]

@mike.alvarez

I can say that workflow improvements, loops and branching are all my radar and being considered for H1 initiatives.

Being “considered for H1 initiatives” honestly raises more questions than it answers.

Can you clarify what that actually means? Is this active development with a realistic delivery window (and if so what rough timeline would that be?), or early planning that could still slip another year or two?

Loops, parallelization, and real if/else branching aren’t experimental features. They’re foundational. They’ve been some of the most requested items for years, they’re table stakes for modern development, and bubblers are already forced into fragile and inefficient workarounds to compensate (e.g. Solving-complex-bubble-problems ).

Given that history and impact, it’s hard to understand why this is still framed as under consideration rather than a committed platform investment. Even a blunt answer on priority and intent would be better than ambiguity.