Monthly Community Update - March 2026

[josh]

Hi everyone,

This is the March community update. You can read February’s update here.

We spent a lot of the last few weeks updating our roadmap in partnership with Brian, our new VP of Product. (If you haven’t read his intro post from November, check it out here — it gives a lot of context for how he thinks about product and what drew him to Bubble.) Today, I’m excited to share it with you in the “looking ahead at Bubble’s roadmap” section below.

The short version is that we’re hyper-focusing on three areas: AI as our primary focus, with mobile and the core editor experience following closely after. In fact, you’ll see those focus areas clearly reflected in the work we completed in February and have already started working on for March.

Let’s get into it.

What we shipped last month

AI

A big one: We upgraded Bubble AI’s app generation capabilities to Claude Sonnet 4.6 the same day Anthropic released the model. This was especially exciting for the team — we got to work with Anthropic to test the model and give feedback before the public release. Sonnet 4.6 brings some distinct style improvements to app generation that make that initial wow moment a lot more compelling. You can read more and see side-by-side comparisons of generations before and after the upgrade here.

Also, the team shipped some improvements to how prompts, caching, and parallel pipeline steps work in the background that made app generation 2x faster than it was before.

We also shipped a number of updates to improve the AI Agent experience: the Agent now anchors to the bottom right of the window, has a refreshed design that’s easier to follow, and renders rich elements in its responses so you can click directly from the chat to the relevant page in your app. We also reduced false positives (when the Agent claims it made a change that it didn’t) and improved how it creates UI components based on your app’s existing styles and layout best practices. See the welcome experience updates here.

If you typically build without Bubble AI — or haven’t tried it in a while — I’d encourage you to give it a go. You can start by building a new app from your homepage and selecting Create with AI.

Mobile

Native mobile apps built on Bubble now run on the newest React Native architecture, which brings improvements to startup time (especially on Android), scrolling, and animations. If you have a live build, you’ll need to submit a new build to the app store(s) to get the upgrade. Read more here.

We also shipped an update that improves image loading time for high-quality images by automatically condensing them to 1080p — about a 60% reduction in load times, especially noticeable on weaker cell service. Read more here.

Editor

Now that the API Connector lives in its own tab, we’re focused on shipping improvements to make it easier to use. This month we started with a hover menu on collection names for quick actions like cut, copy, paste, and renaming; a notes feature for collections and individual calls; and search and drag-and-drop reordering. Read more here.

We’re also continuing improvements to the redesigned property editor based on your beta feedback. The big one: You told us that popovers were feeling like overload, so now only one can be open at a time (with a few exceptions, like the color picker), and they close automatically when you switch tabs or click outside. We’ve also made inputs full-width and stacked for easier scanning, and tooltips now appear after a short hover instead of instantly. Full list of recent changes here, and there’s another update scheduled for this week.

What we’re working on next

AI

• Improvements to Agent response UX: In addition to improving the initial welcome experience, we’re planning to ship updates to how the Agent responds to edit requests to make its responses more legible and interactive. We’re also bringing “undo” functionality directly into the Agent to give you even simpler controls. Look out for those by next month’s update.

• Improvements to dynamic expression generation: Expressions are foundational to how Bubble works — they connect your frontend UI to your data and logic. Better expression generation means better Agent edits across the board. That v2 release is planned for April.

Mobile

• Magic login links: Later this week, we’re planning to launch support for magic login links for mobile apps without requiring a web app. This is a critical feature for enabling app authentication without a password — typically a much better user experience. Before this update, you could build a magic login flow for native mobile users, but it required a web app. Once it goes live, the Send magic login link workflow will give you an option to select a mobile view. Keep your eyes on the forum for that announcement.

• Mobile plugin editor: We’ve been alpha testing the mobile plugin editor and are almost ready to open up access to more developers in beta so they can build features that our mobile editor doesn’t yet support. We’re on track for that to happen in April.

Editor

• Property editor beta: In addition to the popover improvements already released, the property editor team is working on putting dynamic expressions back in-line instead of automatically expanding them into a popover. We’re also improving hierarchy and visual contrast to make the whole thing easier to scan. These updates are released on a rolling basis. As a reminder, you can access the updated property editor in your app settings — if you turned it off on initial release, I’d encourage you to give it another try and let us know what you think.

Looking ahead at Bubble’s roadmap

As I shared in January, we’re going all in on AI. Impressive, seamless AI experiences are table stakes for new users discovering us for the first time, which matters because growing the next generation of successful builders strengthens the entire ecosystem. The more high-quality apps that get built, launched, and scaled on Bubble, the more opportunities there are for agencies, experts, template creators, plugin developers, and power users to thrive.

So our goal is simple: We want that first app creation moment to feel magical. And when builders move into the editor, they shouldn’t stall — they should feel ready to keep going. That means continuing to make Bubble AI and the Agent best-in-class while also investing deeply in mobile and the core editor experience. We are also continuing to work behind-the-scenes on platform reliability and scalability, with a special emphasis on making sure our Dedicated offering is reliable and scales affordably with our customers.

So, here’s what Bubble is focusing on for next few months:

AI

• AI Agent for non-AI apps: A lot of you have asked about getting Agent access in apps that weren’t originally generated with AI, and we’re actively working on it. The first phase is Agent access in new blank apps and templated apps. Rolling it out to existing apps is the next piece of that puzzle — we’ll be working closely with select Bubble Ambassadors and Gold tier agencies to begin testing by April.

• AI Agent compound edits: We’re working on giving the Agent the ability to make multiple changes at once. This is more in line with how people actually interact with the Agent — you want to prompt full features, not piece together UI, then data, then logic separately. We anticipate that release will be out in May.

• Self-correcting AI Agent: Right now the Agent doesn’t do a great job of checking its own work, which limits how helpful it is for complex tasks. We’re evolving the Agent’s reasoning and observation capabilities so that it’s better at recovering from mistakes and implementing what you’re actually trying to build. Part of this initiative also includes integrating access to the issue checker. That work is slated to be complete by June.

• Remixes: Start from a templated base app, then choose your customizations before generating. We’re aiming to have the first version of that feature out by May.

Mobile

• Deep linking via mobile push notifications: Besides in-app purchases and the mobile plugin editor, this is our most requested mobile feature. Deep linking lets you direct your user to a specific view in your app, rather than just opening to the home screen. This work will add a Navigate to input to the existing push notification workflow. We’re targeting April for that.

• Extending Agent capabilities to mobile apps: Right now you can use Bubble AI to generate mobile frontends, but we want the Agent experience for web and mobile to reach feature parity. We’re working on giving the Agent the ability to edit UI, workflows, data, and more on mobile. Expect updates on a rolling basis, similar to how we’ve been shipping Agent capabilities for web, into late spring and early summer.

• Automatic keystore file generation: Google requires you to generate a keystore file during the app connection process, and right now the only way to do that is through your terminal — a pretty technical ask. Automating that on your behalf should make mobile builds and submissions significantly easier. We’re aiming for early summer for that release.

Editor

• Native JSON support: Right now, when you connect to LLMs via API calls and return structured data in JSON format, it can be hard to use that data in Bubble workflows. We’re hoping to ship a solution that makes it easier to parse JSON natively sometime in June.

• In-context API and dynamic expression generation: We’re still shaping the exact experience, but the goal is to create entry points that make it easier to generate API calls and write dynamic expressions in context in the editor — two big stumbling blocks for less technical users building agents and other AI-powered apps. We expect that to be in your hands by the start of summer.

If you have questions about this or anything else I’ve shared today, you can bring them to Emmanuel’s next live AMA on Wednesday, March 11 at 12 PM ET — he’s excited to chat about the roadmap with you all there.

In case you missed it

Don’t miss the latest episode of our podcast, The New Build. We talked to Vai Singh, who built Turing App on Bubble after growing his YouTube audience to 50K+ subscribers. Subscribe on YouTube, Apple, or Spotify — the next episode is coming out tomorrow!

And check out these two spotlights come from Launch Lab — our program for builders who are close to launching and want hands-on support getting there:

• Ten Points was built by a school leader and an enterprise tech guy — neither of whom wrote a line of code. They iterated directly with partner schools to build a platform that brings behavior recognition and wellbeing support into one connected daily rhythm, now used by hundreds of students every day.

• Pawshare Club is a community-powered pet care network connecting local owners through shared sitting, walking, and playdates, with a reward points system that keeps exchanges fair. A bootstrapped dog owner in Australia built the support network she wished had existed — 377 users and real transactional revenue since launching in August 2025.

Built something on Bubble you’re proud of? Share your story with us — we’re always looking for builders to feature.

New hires

We had four new people join the team this month, mostly on the engineering side. Welcome to Emre is a Senior Engineer on the platform team (same as Glenn and Simon); Glenn and Simon, senior software engineers on Platform; and Rain, a technical success specialist.

If you’re in the job market, make sure you check out our open roles — there are a lot of them!

That’s all for this month.

— Josh and Emmanuel

[johnny]

@josh

Native JSON support: Right now, when you connect to LLMs via API calls and return structured data in JSON format, it can be hard to use that data in Bubble workflows. We’re hoping to ship a solution that makes it easier to parse JSON natively sometime in June.

Nice!

Great updates - lots of look forward to!

[Sergiofojr]

Hi there,

First of all, I truly appreciate the transparency and the level of ambition behind the platform’s evolution. It’s clear that a lot of thought and effort is going into pushing the boundaries — especially around AI — and that’s exciting to see.

That said, I’d like to respectfully share a concern from the perspective of someone building serious production products on Bubble.

The current state of Mobile still feels significantly behind what many of us need to confidently scale real-world applications. It has been in Beta for quite a long time, and while we understand that innovation takes time, the lack of visible, consistent progress — particularly around performance, stability, and core optimization — is becoming difficult for teams that depend on it.

In several support cases, the common pattern has been either suggested workarounds or “please wait for the team,” which often translates into months of uncertainty. For mission-critical products, that creates real operational risk.

AI is undeniably important and strategically forward-looking. However, from a product foundation standpoint, Mobile still requires substantial refinement. Many of us would welcome a clearer roadmap, stronger prioritization, and more predictable iteration cycles on Mobile performance and reliability.

Bubble has always stood out for empowering builders. To keep that trust strong, the Mobile experience needs to evolve from “promising beta” to “production-ready foundation.”

Thank you for listening — and for continuing to build something ambitious. I share this feedback because I care deeply about the platform’s long-term success.

[J805]

@josh

a notes feature for collections and individual calls

Thank you! Looking forward to this. This is such a little thing but a HUGE help for documentation purposes.

[randomanon]

Hi @josh, thank you for the update.

@josh

Remixes: Start from a templated base app, then choose your customizations before generating. We’re aiming to have the first version of that feature out by May.

This is a very smart direction. I’ve recently spoken about how Components are currently neglected despite being one of Bubble’s strongest potential hooks for new users. With this, you get the best of both worlds (Components + AI). Convex even has something like this for workflows. I think it’s a wasted opportunity to not have “Remixes” for Backend Workflows (e.g., AI agent, Stripe integration, etc.) that could DRASTICALLY reduce the build time for new users. This feature alone might substantially reduce new user churn.

I think your internal “Bubble” builders should prioritize this over building internal tools. The “Bubble.io is built with Bubble” meme is kind of fun and cutesy but not really an efficient use of your team’s time compared to, “build out a bunch of common workflow that most apps would benefit from incorporating.”

@josh

We’re also bringing “undo” functionality directly into the Agent to give you even simpler controls.

Great, it’s extremely important to have a robust system of “diffs.” I don’t know what’s happening with the “new Changelog” feature that only certain tiers have access to but when it comes to the agent, it needs to be as robust as Github branches. Hopefully this undo works better than the current one.

@josh

That v2 release is planned for April.

@josh

begin testing by April.

@josh

We anticipate that release will be out in May.

@josh

That work is slated to be complete by June.

@josh

We’re hoping to ship a solution that makes it easier to parse JSON natively sometime in June.

@josh

We expect that to be in your hands by the start of summer.

OK, just understand that right now in the software world, companies are shipping in dog years. So one human month is equivalent to eight “dog months.”

@josh

As I shared in January, we’re going all in on AI.

Good, but is this a case of too little too late? I feel that there was a certain inflection point that we’ve already passed. With the influx of memetic slop agent builders like “OpenClaw” and its ilk, is there any mindshare left for Bubble? Even if Bubble is rationally the better option for a new builder, does this actually matter if it’s not salient in the attention marketplace?

Strangely there is a certain community tool built by 1 person (which you’re no doubt already aware of) that is somehow advancing at a faster rate than a VC-funded company with dozens of engineers. Yes, you have a dozen more surfaces to work with, yes, there are formal SDLC processes at companies, etc. etc. But it just seems like everyone is shipping faster than Bubble right now. What’s going on?

[JamesC]

Thanks @josh for the call out of Ten Points! Super happy to be part of Launch Lab

[tomvigneres]

@josh

Native JSON support: Right now, when you connect to LLMs via API calls and return structured data in JSON format, it can be hard to use that data in Bubble workflows. We’re hoping to ship a solution that makes it easier to parse JSON natively sometime in June.

Really bullish on this! we definitely need native JSON support in Bubble

[ihsanzainal84]

@josh

Native JSON support: Right now, when you connect to LLMs via API calls and return structured data in JSON format, it can be hard to use that data in Bubble workflows. We’re hoping to ship a solution that makes it easier to parse JSON natively sometime in June.

So so hyped for this!

[josh]

@Sergiofojr

The current state of Mobile still feels significantly behind what many of us need to confidently scale real-world applications. It has been in Beta for quite a long time, and while we understand that innovation takes time, the lack of visible, consistent progress — particularly around performance, stability, and core optimization — is becoming difficult for teams that depend on it.

Thank you for the feedback – we understand and we are doing a lot of work here. We don’t usually mention bug fixes in this update, but you can see the list here: Product Updates and Release Notes | Bubble

The architecture release mentioned above is also a big deal – we’re still sorting out a handful for follow-on issues, but this puts us in a much better position moving forward.

@randomanon

Strangely there is a certain community tool built by 1 person (which you’re no doubt already aware of) that is somehow advancing at a faster rate than a VC-funded company with dozens of engineers. Yes, you have a dozen more surfaces to work with, yes, there are formal SDLC processes at companies, etc. etc. But it just seems like everyone is shipping faster than Bubble right now. What’s going on?

No need to be coy – we love Buildprint, and have been chatting with @georgecollier about getting it in front of more users.

We’re also aware industry development pace has been accelerating tremendously because of AI, and we’re doing a lot behind-the-scenes to speed things up ourselves… the dates we’re sharing are dates we feel at least fairly confident about hitting, since we’ve overpromised in the past.

I would say, though, that the meme that Bubble isn’t shipping fast is just a meme. I think it was true a year or two ago, but if you look at our total output across all parts of the product, a lot is going out the door each month. And I’d point out that v1 of Buildprint was released last April (so presumably dev worked started before that): this is a space that George has been playing in for an extended period of time.

I don’t want to overstate the case – there’s plenty we can do to speed things up, and driving faster/better engineering execution is what I wake up every day thinking about (when I’m not thinking about AI) – but I do think the forum discourse on this has built on itself a bit into a place that I’m not sure is accurate

[randomanon]

@josh

We’re also aware industry development pace has been accelerating tremendously because of AI, and we’re doing a lot behind-the-scenes to speed things up ourselves…

Awesome!

@josh

the dates we’re sharing are dates we feel at least fairly confident about hitting, since we’ve overpromised in the past.

Fair enough, but let me contextualize the date issue a bit further down.

@josh

I would say, though, that the meme that Bubble isn’t shipping fast is just a meme. I think it was true a year or two ago, but if you look at our total output across all parts of the product, a lot is going out the door each month.

All the R&D and engineering that went into native mobile + the new workflow editor created a huge clog in the system. Now I feel like stuff is happening again, yes. You actually said it perfectly. There’s a high “total output across all parts of the product.” But I would say that the “effective output” (factoring in the “wind chill factor”) is much lower. It’s scattered across various systems/surfaces rather than being focused on delivering “magic.” I look at the updates, I think to myself, “ah, that’s kind of neat.” I have not felt any significant “enablement.” The core experience of the editor still feels archaic and 1990s-esque for various aspects of Bubble development. When you’re forced to do manual labor because Find & Replace (in an app-wide context) doesn’t exist, it just feels like a waste of time. Sometimes I legitimately feel like a Boomer who is hunt-and-peck typing their way through app development (no shade to Boomers). The AI Agent actually obviates the need to fix various platform deficiencies through sheer brute force (e.g., “Go and replace all expressions with this.”). This is a major MAJOR boon to your company as it handwaves entire complex systems that no longer need to be developed (e.g., proper expression composer).

Which leads me to why I mentioned the dates in a nitpicky and critical way. From where I’m sitting, there hasn’t been any real progress on the AI Agent. This video came out 11 months ago, but it’s not real yet. You could argue the models themselves weren’t where they needed to be at the time, and while that was true back then, it hasn’t been the case for while now.

AI Visual Development: The Future of Bubble AI - YouTube

Yes, the “new app generator” is marginally better, but without hitting the critical threshold of usefulness. And for the overwhelming majority of current Bubble builders, there’s very little overlap between that product and their daily workflow. Mind you, in this timespan, other companies have straight up built their version of “Bubble” (Cursor Visual Editor), as I predicted would happen 1 year ago:

@randomanon

Let me tell you what the real issue is. You have three categories of products, frontend platforms (weweb, toddle, etc.), backend platforms (xano, supabase, etc.), and what I charitably refer to as “Slop Builders” (Bolt, Loveable, etc.). There is a secret fourth category which is actually just one product (replit). They may not be actively aware of it right now but all of them will eventually converge to the same point (Bubble). You guys had the early vision figured out a long time before anyone else, and you had a 10 year head start. But they are rapidly closing in due to AI. They are coming for your “Day 2.”

So to contextualize my earlier comments about shipping speed: My concern is that they’re going to build you before you build them. I think you’re at an existential competitive inflection point which will require you to throw caution to the wind and hyperscale. However, I fear that the old “bootstrapped mentality” of “sustainable growth” will get in the way of this. I hope I’m wrong.

Thank you again for taking the time to think through and write out a response. I really do believe that Bubble’s core platform + abstraction layer + the AI Agent is objectively unbeatable as a product if executed correctly. There’s a lot to look forward to!

[senecadatabase]

Right now, Bubble’s AI is behind. It needs and should be moving at a greater pace.

I believe they can, I’m just not sure what’s going on with the slow progress.

We have Bubble user’s that are able to surpass it very quickly, as we’ve seen recently with @georgecollier etc.

It’s almost unacceptable that things are moving this slowly on Bubble’s side.

I’m holding out hope, but it would be nice to see things that offer more hope

[asked111]

Good update thank you and all sounds positive.

I don’t like being negative and prefer to be constructive, so I hope you take this as that …

I urge you to remember/realise that even with ai working perfectly, you will struggle to compete if you don’t get the absolute fundamentals nailed - page load & database speed, and SEO.

Because ultimately all of the web builders from the most beginner friendly (like Wix) to advanced platforms will all eventually converge around AI + visual editing. They will just offer slightly different take on how they do it. So to keep bubble toe-to-toe you simply will have to nail the essential basics.

I am currently going down a path of using static Cloudflare hosted index & SEO pages and an external database satalite for speed reasons, but I REALLY wish I didn’t have to even consider this . As some have said before, that is a slippery slope towards rebuilding everything outside of bubble.

Something to think about but I have faith you already are

[boston85719]

@josh

No need to be coy – we love Buildprint, and have been chatting with @georgecollier about getting it in front of more users.

@josh can you confirm that a service like buildprint is not in violation of Bubble acceptable use policy?

Areas of the policy that 12 months ago gave pause for building such services include the below:

A. You may not make unauthorized copies, modify, adapt, translate, reverse engineer, disassemble, decompile or create any derivative works of the Platform or any content included therein, including any files, documents, postings, or documentation provided by Bubble or any other user of the Platform (or any portion thereof).

B. You may not determine or attempt to determine any source code, algorithms, methods or techniques embodied by the Platform

G. You may not engage in “framing,” “mirroring,” or otherwise simulating the appearance or function of the Platform.

I know that there have been multiple component library services that run through google chrome extensions over the years, which apparently necessitate the usage of the Bubble editor and for it to be open while interacting with the chrome extension for purpose of using the current user session and browser copy/paste.

The a service like buildprint works for making in app edits I assume does not require the bubble editor to be open in the browser, and is using a server to run these processes through the bubble editor APIs.

There have been a couple of other services, like a logs service (not buildprint - one launched a day prior) that also are likely using bubble editor APIs to provide their service.

Can you please confirm if use of editor APIs is not against the acceptable use policy and that we can use our own user session tokens in server actions and do not need to have the editor open in a browser to leverage an existing user session token to run such services? I personally can not tell if I am reading into those acceptable usage policies correctly or not.

It would only help improve the platform ecosystem to grow and expand to see more products/services similar to buildprint, so having a clear understanding of whether or not they are truly permissible would be great to give others clarity on what Bubble finds acceptable and not acceptable at this time.

[georgecollier]

@boston85719

@josh can you confirm that a service like buildprint is not in violation of Bubble acceptable use policy

not Josh but

@georgecollier

As long as we’re polite to the Bubble platform’s systems, and positively benefitting the community, Bubble has no interest in stopping community built tools. Just before I launched NQU Secure (shortly after Flusk acquisition), Josh himself told me he views third party tools as a healthy part of any successful ecosystem.

So like, Bubble probably has the right to say no, but as long as you’re not a dick, why would they want to?

[randomanon]

@boston85719

@josh can you confirm that a service like buildprint is not in violation of Bubble acceptable use policy?

He just effectively confirmed it here.

@josh

we love Buildprint, and have been chatting with @georgecollier about getting it in front of more users.

Not sure if you’ve ever worked at a corporation before but the purpose of the Terms of Service is solely to protect the company’s business interests and users. Just because there’s broad language in there doesn’t mean every violation will always result in legal action. If Buildprint (or any other tool) presented a threat to Bubble’s business interests or users, then they would take legal action.

[fede.bubble]

Hey all, moving most of the convo around Buildprint access, legal, compliance, etc to a new topic so we can keep this topic discussion around the actual announcement and what it covers.

You can find it here:

Terms of service concerns around building third party tools

Yes, I read that commentary on the other thread previously. What I am hoping for is @josh or @emmanuel or perhaps @fede.bubble to provide an official bubble stance on the use of the APIs for these types of products/services so it is clear from the company what is an acceptable use versus not. My main concern, is whether or not we can use a server to trigger these API calls or not. I do not want to have to build a product into a Google Chrome extension as that is adding an extra layer to my pro…

[code-escapee]

@josh

I would say, though, that the meme that Bubble isn’t shipping fast is just a meme. I think it was true a year or two ago, but if you look at our total output across all parts of the product, a lot is going out the door each month. And I’d point out that v1 of Buildprint was released last April (so presumably dev worked started before that): this is a space that George has been playing in for an extended period of time.

---

Come on Josh. You’re technically right, but you’re missing why people keep saying this.

The Buildprint point deserves a direct response. V1 and v2 are not the same product. V1 was about turning Bubble’s JSON into human-readable documentation and making it queryable. V2 is a new tool entirely. An autonomous agent that diagnoses and fixes real builder issues. It was built in a few weeks by someone with a full-time job whose livelihood does not depend on this. He built it because he doesn’t want to be forced off Bubble. Bubble has had a funded team working on this problem for two years and hasn’t shipped anything close.

That’s where the meme comes from. Not “Bubble ships nothing.” It’s that Bubble ships, but not the things we’ve been asking for. And when it does ship something in that direction, it’s a watered-down, hype-first version that doesn’t deliver the actual utility we need. The AI Agent is the clearest example of that. Years of anticipation, and what we got doesn’t work on our existing apps, hallucinates completed changes, and has broken bindings in its own announcement video.

Then there’s the other category: things that haven’t shipped at all, in any form, after years of asking.

• Front-end to back-end communication that feels native, not duct tape
• Actual looping primitives for workflows (not “schedule API workflow on a list” as the answer to everything)
• Parallelism and fan-out/fan-in patterns that don’t require hacks and fragile orchestration
• Native UI elements so we stop rebuilding basic platform controls from scratch

And it’s not just Buildprint. I built the first version of BUBS, my Bubble AI agent, in a single day. It wires front-end UI elements, reads the Issues Drawer, fixes them, and works across all Bubble apps including existing production apps. Improvements since then have come from actually using it. That’s the bar. Bubble’s AI Agent launched in October for AI-generated apps only, and five months later it still isn’t usable as a daily driver for apps we actually have in production.

---

Here’s what Bubble shows you the agent can do:

Here’s what actually happens:

“Changes approved and applied.” Twice. 28 errors before, 28 errors after. Then when you push further, it tells you it can’t read the issue drawer and can’t accept a screenshot and can’t fix the issues when pasted as text..

And it’s not just user sessions. In a new AI Agent video, Bubble’s own produced content has fields showing [not found: destination_text] and [not found: trip_notes_text]. Broken data bindings baked into the marketing material.

---

So yeah, maybe the company is shipping a lot across surfaces. But when someone with a full-time job and zero financial stake can ship a leap like Buildprint v2 in a few weeks while the core builder asks sit for years, “we ship a lot” does not land as an answer.

The question isn’t whether Bubble is shipping. It’s why the things we’ve been asking for the longest are either still untouched, or show up as a press release before they’re ready to actually use.

[agiledood]

Sorry, the AI improvements are still comically bad. I made a new app from scratch for a simple download marketplace. It was created quickly, looked great, the test data and list/detail stuff worked great but ZERO followup prompts worked:

• create me an option set… (AI Agent: Sorry I can’t do that, here’s how to do it yourself.) - No big deal…I created it.
• remove the ‘content_type’ field from the uploads table and add a field ‘type’ using the option set ‘type” (thinking thinking thinking…all done, but one update couldn’t be applied’) - actually none of them could be.
• make me a popup on this page so i can add records to the upload table (thinking thinking…it made a new table called uploads, said it would now add my new option set ‘type’ and replace the existing content_type field with the new ‘type’ field. It did not. It added a blank field on a new table I didn’t ask it to create. And it didn’t make the popup, or mention it.

To be fair, I referred to the table as ‘uploads’ which didn’t exist but the agent should be smart enough to know that if I request an UPDATE to a table that doesn’t exist, the agent should say “hey dummy, there’s no table with that name, but there IS an upload table!” - not create a new table.

Guys, seriously considering ditching the agent or at least make an agent that can design forms perfectly. I’m guessing everyone with a reasonably large site has built their own basic CMS and that’s the most irritating part. “make me a data entry form for the ” THAT is useful.

I still absolutely love the platform, and I am completely reliant on it now given how big my site is. Maybe it’s time to consider a reorg and splitting the platform into ‘ai agent crap for newbies who want to build mvps and one page simple sites’ and another for power users that use it as a serious development platform. It worked for Active Campaign when the spun up Postmark.

[bonyeckj]

Finally native json parsing! Thanks!