Hey folks! Not Quite Unicorns works on some of the most complex Bubble apps out there. You might remember that almost a year ago, we released Buildprint 1.0 - the documentation platform for Bubble apps that lets you chat with your app. Today, we’re releasing Buildprint 2.0 in alpha (the Bubble forum gets it first :smiley: ) Want to try it out? Connect your app for free @ https://getbuildprints.com :tada: Buildprint 2.0 is a complete rebuild. Old Buildprint is no longer accessible. It’s now an AI and observability platform that connects your Bubble app to any AI - Claude Code, Cursor, ChatGPT, or any MCP-compatible client - and gives you the advanced logs, alerts, and debugging tools that Bubble doesn’t. Buildprint also lets you connect your Claude/ChatGPT subscription and chat right within the web UI to get started easily. Here’s what it does. - Use any AI with your Bubble app - Advanced logs and observability - Workflow log trace explorer - Alerts and anomaly detection - Automatically triage and debug issues with real data - Plan new features and identify optimisation opportunities - Get diffs and understand changes between branches The goal is to get people excited about Bubble again and augment Bubble with the kinds of features that scaling apps and teams need to keep growing quickly. Use any AI with your Bubble app Buildprint connects your AI to your Bubble app. This means your AI has full context - your database, workflows, logs, and app structure - so it can give you accurate answers instead of generic ones. Two ways to use it: In your coding agent - Generate an MCP server URL and add it to Claude Code, Cursor, or Codex. Ask questions like “Why are users getting blocked at checkout?” and get answers grounded in your actual app data, logic, and logs Directly in Buildprint - Connect your ChatGPT or Claude subscription and chat right in Buildprint. No setup beyond linking your account. [image] [image] Advanced logs and observability Bubble’s native logs viewer is limited. You can’t run complex queries, you can’t aggregate data, and you can’t see what’s actually happening at scale. Buildprint streams your logs to a purpose-built analytics backend, giving you: Real-time dashboards – workflow runs, API calls, error rates, unique users, all updating live Advanced querying – filter by time range, user, workflow, log message, or run full advanced queries with Buildprint AI for complex analysis. Logs explorer - view logs, workflows, and traces in detail to accurately follow logic through your application. [image] [image] Alerts and anomaly detection Set up monitors that watch your app and notify you when something goes wrong – before your users tell you. Three types of monitors: Threshold – alert when a metric crosses a value you define (e.g. error rate > 5%, workload for a given user > 5,000 in an hour) Event matching – alert on specific event patterns (e.g. a particular workflow fails) Alerts are delivered via email to your team by default, or you can enable webhook forwarding for Slack, Discord, or any other tool. [image] Debug any issue This is where Buildprint changes how you work. Instead of manually trawling through Bubble’s logs trying to figure out what happened, you can just ask: “What caused the ‘order_status’ field to change to ‘cancelled’ for user X?” "Why does user jane@example.com see a blank page when they load the dashboard?" “Show me all errors from the checkout workflow in the last 24 hours” The AI has access to your real data and logs, so it can trace through exactly what happened – which workflow ran, what conditions were met, what data changed, and why. [image] Plan new features and identify opportunities Buildprint understands your app’s structure – the database schema, workflows, pages, and how they connect. So when you’re planning a new feature, you can ask: “What tables and fields power the onboarding flow?” “Which backend workflows are unused and save to remove?” “How does the payment workflow connect to the user’s subscription status?” “What would I need to change to add a referral system?” You get answers that are specific to your app, not generic Bubble advice. Diffs and code review for teams For teams that need to understand what’s changed in their app, Buildprint tracks versions and lets you compare them. This is particularly useful when: Multiple developers are working on the same app You need to review what changed before deploying to live Something broke and you need to identify which change caused it To use it, ask Buildprint to explain or investigate changes in two versions. [image] How it works Sign up for free at getbuildprints.com Connect your Bubble app and verify ownership Choose your setup – use Buildprint’s built-in chat, or plug the MCP server into your preferred AI client Optionally enable advanced logging – stream your Bubble logs for real-time analytics and alerting Setup takes a few minutes. What’s coming soon? Buildprint 2.0 is in alpha. We’re actively developing it and would love feedback from the community. If you’re working on a complex Bubble app and are frustrated with the current debugging and AI tooling, give it a try and let us know what you think. Over the next few weeks, we’ll be adding: PR like interface for code reviews, understanding changes, and requesting merges integrations with tools like Linear so you can associate Bubble branches with specific issues and keep them in sync automations to programatically call Buildprint (e.g when support ticket received → kick off Buildprint agent to investigate and triage) Agentic editing And, most importantly, Bubble app editing. We are not going for an agent that can vibe code an app to 80%. Instead, we’re working on the agent that lets you take your existing app from 80% to beyond. That is the holy grail, and is something I look forward to sharing with you all very soon. Happy to answer any questions below.

[image] georgecollier: Well, from what I can tell, they already do - it’s just soft-disabled. But it works, you just can’t select the type from the params dropdown. If I had to guess, it’s for performance reasons of some sort. Very interesting. I wonder what other “weird” unsupported things might AI editing insert into a Bubble app unbeknownst to us and cause strange problems down the line? There’s an argument to be made that even the visual UX editor has things like this (e.g. certain conditions can only be copied from elements into workflows and they work fine) but AI editing opens up a whole new set of unknowns.

Ya, when we find stuff like this we improve Buildprint to stop it happening in future as it may not be officially supported :slight_smile:

I do wonder about things that might not be immediately apparent through the editor. If something can slip through the linter and the issue checker and just silently sit there for a while… Kind of spooky to think about. I know it’s an edge case (and unlikely one at that) but I’m a tail risk kind of guy. On the flip side, as models continue to improve and you continue to add to the system prompt + your programmatic guardrails, I guess the product itself can scan for and fix these types of issues as effectively as it can create them.

The engine which processes app JSON into code is entirely client side, so it’s easy to go through and identify problematic areas. But for sure there’ll always be stuff to improve and I can’t offer any guarantees with Buildprint yet.

Buildprint organised 345 workflows on this page into folders in one shot based on which section of the page the workflows belong to. Agent spent about 20 minutes doing it. It didn’t even take a shortcut like grouping them by event type or action or something. They’re genuinely semantically organised! [IMG_1416] [IMG_1415]

Yeah good example! Did you do this with Buildprint tho :thinking:

No hahaha, I just created the expression in a group element (with the final part as “Current User”), then copied it to the privacy rules and changed the final part to “This user”.

I was gonna say, we already check for stuff like this :joy: Don’t scare me like that

I haven’t tested if it actually works, but it doesn’t seem to be flagged as an issue.

Announcing Buildprint AI: Build on Bubble with Claude Code and Codex 🎉

Showcase

Newed February 27, 2026, 2:22pm 114

Maybe something like that.

A simple example of list operators not currently supported in privacy rules, but that can be enforced.

"ILLEGAL EXPRESSIONS" that AI Agents are managing to write in Bubble

Topic		Replies	Views
🎉 Introducing Buildprint \| Auto-generated documentation for complex Bubble apps Showcase	166	4821	November 27, 2025
Monthly Community Update - March 2026 Announcements	21	2841	March 11, 2026
Our engineering workflow before vs after using BuildPrint (Bubble agentic engineering platform) Tips	16	759	July 5, 2026
Build/Manage/Update Bubble.IO App via Claude Code Questions	7	348	April 28, 2026
Why it's so quiet here? We got Global Expressions, data return from Custom Backend WF's and DARK MODE! Implemented	14	644	June 5, 2026

Announcing Buildprint AI: Build on Bubble with Claude Code and Codex 🎉

Related topics