Romain

Certified @Bubble developer, up in the mountain! Turning caffeine into apps and mistakes into lessons

[Romain @_rmnddr]

This @bubble cap is actually quite practical after all…

Happy bubbling y'all !

[Romain @_rmnddr]

FINALLY 🙌

[Bubble @bubble]

Native JSON parsing is here — if you're building with an LLM API, the structured data coming back from your calls is now directly accessible in Bubble without a single plugin or workaround.

[Romain @_rmnddr]

@fede_bubble @bubble Makes sense for the 1M case. The painful part is the silent mismatch: Search:count = 9, iterated list = 4, no warning. Privacy isn't it, `:items until #N` fixes it without changing auth. Would really help to have docs on the default cap, or a warning when count ≠ returned items

[Romain @_rmnddr]

Hey @bubble, spent a 2 hours debugging a marketplace bug. A Search returns 9 matches (Search:count = 9), but when passed to a custom element as a param, only 4 reach the receiving RG. Silent lazy-load. Fix: add `:items until #N` to force materialization. Why is this the default?

[Romain @_rmnddr]

@bubble I mean, its a list of 9 items ...

[Romain @_rmnddr]

@george_nqu @bubble @ClausISaidSo Give this man a chair on bubble's board for @bubble 's sake ...

[George Collier @george_nqu]

Everyone I work with that is thinking of leaving @bubble starts with ‘I love Bubble, but…’

Which is super interesting. Because how could a company with such positive user sentiment fumble their lead so catastrophically. What is Bubble doing that makes advocates consider leaving?

The answer is probably more in what is Bubble *not* doing.

If you look at the feature releases for the last 6-12 months, the biggest feature released by Bubble for existing serious users with real businesses on Bubble that don’t have native mobile apps, is button icons. Button icons! In 2026!

And the most important ‘feature’ released altogether for this user segment has been Buildprint, which isn’t even built by Bubble!

The property editor updates are fine but don’t move the needle for anyone because experienced users already knew the ins and outs of the old ones so updating it would never be a priority for them.

The biggest question we have as users is ‘who is Bubble for?’. If you’re non-technical, you can’t really build anything substantial on Bubble and claiming as such is just wishful thinking. If you’re technical enough to use Bubble, you’re also technical enough to use AI-assisted development with code.

Not even the Bubble TEAM knows who Bubble is for. @emmanuel_s has this even been agreed? How can a product roadmap exist if you don’t know who you’re building for? ‘Win first builds’ but win first builds from *who*? There’s no sense of direction or coherent vision that we can see externally based on Bubble’s releases and marketing.

The serious users of the platform have basically been given nothing over the last year, so why would they believe they should stick with Bubble and trust they’ll get the features they need, when Bubble’s offered no words or actions to that effect?

For many it will be too little too late and the decision has been made to move off of Bubble. But for most, who are still unsure, Bubble has the opportunity to extend an offer to them and commit to features that they need, like workflow branching, returning data from backend workflows etc.

But on the current course, it will keep being ‘next year we’ll work on power users’ but it’ll never be attended to because the new user experience will always take priority.

It’s all good prioritising new users, until the existing users spending thousands of dollars a month decide to leave and revenue starts leaking from the back door of the platform.

I’m not naive to the importance of continued growth and staying relevant in an AI world but the reality is that a significant number of apps are now considering migrating (many for the wrong reasons, and they don’t actually need to migrate - but that doesn’t change the maths for Bubble).

Bubble has an opportunity over the next few months to extend an olive branch to serious apps and give them things they’ve been waiting for years for. Waiting any longer will be too little too late for many.

I’m rooting for Bubble and invest in the platform because it can have a future if it moves in the right (or even just any) direction.

I’m not going to just stand by it if we get nothing back, though.

[Romain @_rmnddr]

@george_nqu @bubble Sweet, thank you!

[Romain @_rmnddr]

@george_nqu @bubble Ok it's telling me I don't have Buildprint credit enough to deploy an agent, is it normal ?

[Romain @_rmnddr]

Hey @bubble why can't we just have a "is empty/ is not empty" for a boolean? Especially if it's a Backend WF or a custom event parameter???

[Romain @_rmnddr]

@george_nqu Obviously it’s not 100% the same because it used my styles, which is great!!

[Romain @_rmnddr]

I tried @george_nqu Buildprint building feature again today and maaaan! This thing has become great. I just told him, “Replicate this, I created an RE with the following ID,” and he did it. It even understood I made a custom checkbox as another RE, used it, and filled properties!

[Romain @_rmnddr]

@ToolsOttho Comment ça ils prennent des pauses ? AU CHARBON LÀ!

[Romain @_rmnddr]

@victor_bigfield Nice! Trying to quit smoking here and went for 10km in 2 days, it's a start! Nice reading you, it's motivating!

[Romain @_rmnddr]

@johnnylinsf @bubble @LemonWay Now I find myself thinking "How can I find a way, AI can do that for me ?" instead of diving into the problem. But I guess I'm more efficient this way! But I feel better with your point of view!

[Romain @_rmnddr]

I talk to Claude Code with Buildprint MCP to monitor my @bubble app. Needed @LemonWay payment logs, connected n8n MCP, asked Claude to build a Lemonway MCP. Now I just ask and get answers. Is this the future? Be lazy, code an MCP, talk to Claude?

[Romain @_rmnddr]

@george_nqu @bubble whooops, I did not see that notification, will have a look!

[Romain @_rmnddr]

@george_nqu @bubble 🙌🙌🙌

[Romain @_rmnddr]

@jamesdevonport Exactly the same ! Now its my best friend