Announcing Buildprint AI: The agentic engineering and observability platform for Bubble 🎉

[jsterling]

Custom sounds for threads? Dude, ringtones are so 2000s

Awesome that we can use OpenRouter now!

[parthapatel0723]

I used to get annoyed about netflix being too much…now I pay tons more for every AI subscription

[georgecollier]

you can save $20 by cancelling gemini

[parthapatel0723]

Unfortunately, I’m a Antigravity IDE power user

[building4stripe]

@georgecollier Been stuck at “Checked sync status” for about 35 minutes when using Review. Known issue?

[devin.fraze]

Out of curiosity in hopes I might learn something. Why rename them in the schema you used?

[parthapatel0723]

I have syntax I use throughout my app for various things, it helps me clearly understand what is what, where, and purpose. I have a pretty complex app with 60 or 70 reusables, 25+ data types, and 50+ option sets. This structure helps me break down what data type part this reusable is relevant to in {} so for ex. {Global} represents used universally but {Payment} is all reusables relevant to payment. Then “:” breaks it from the data type side, and provides specific meta such as “Component” (meaning it’s a reused UI element (borrowing modern js framework naming). The last part in () is the element type like popup, group focus or group. I use this syntax: {Data Type} (optionSet) (Element Type). Also, probably my undiagnosed OCD lol. An example from my app:

[georgecollier]

Buildprint MCP was down about 50 minutes ago, should be resolved now!

[randomanon]

@georgecollier

In my experience, the Gemini models aren’t particularly strong and can feel a bit slow because they overthink

I hate Gemini for literally everything outside of “what’s the best way to build this in Bubble” (especially if it’s a hybrid of JS + native Bubble).What do you find best? GPT 5.4?

[produccion1]

in my experience gemini allways has gived me better results on clean UIs, so i will be probably ussing gemini only for UI polishing.

[produccion1]

every time i try to use it: Agent error: M is not defined

[georgecollier]

Should be fixed if you start a new chat

Also don’t use reasoning with the Gemini models, it’s deathly slow!

[senecadatabase]

@randomanon

I hate Gemini for literally everything outside of “what’s the best way to build this in Bubble”

I found Gemini to use more corporate-type language when you ask it to help with an article, etc.

I mean, if you’re writing a white paper for Harvard, it would be ok…

but not talking to the average user.

I always try to write like I’m having a one-on-one with the user. ChatGPT does a pretty good job of writing, usually. Normally, I’ll write something out and then ask for punctuation checks or word usage

[georgecollier]

Buildprint editing comes out of beta on April 20th

Exciting news! Buildprint editing will be released to everyone on April 20th.

You can join us for a webinar with @lucas_bennington and @theo.goldberg to see how Lucas’ Codi Tech (a Gold tier agency) is using Buildprint to deliver work, and how you can deploy Buildprint in your agency, production app, or internal team.

Deploying Agentic Engineering in Your Bubble App · Luma

[lucas_bennington]

Exciting stuff!

[mo.salama]

@georgecollier can we mention Skills while using the agent API?

[georgecollier]

@mo.salama

can we mention Skills while using the agent API?

Not right now I’m afraid

[georgecollier]

FYI we had a regression with expression editing that led to general badness which is being resolved today

[randomanon]

George Buildprint defeats General Badness

[ijmatthews]

Hi @georgecollier - If there is any way to be added to the beta for the editing functionality ? I’ve got a number of changes to make and I’ve debated just halting bubble dev and setting up a full stack app using Claude instead - BUT the AI functionality you’ve developed is game changing and I’d love to properly try it out to see how it goes