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

Have you tried putting them into folders, which let you organise them easily?

Have you tried opening a side chat?

Yes

…True : I’m bad at managing folders, my keypoint is to make sure i don’t loose my functional documentation that i try to contain in 1 conversation and still if you have 20 conversations in your folder, the 1 important is easily lost (it’s coming back to shared documentation between conversation)

Unless i miss something, side chat is not compatible with the buildprint iOS app (testflight)… so i admit i’ve stopped using it

Thanks for coming to office hours everyone! Recording will be uploaded soon.

For anyone that wasn’t there, if you want to be moved off the logs waitlist, upgrade to Bubble’s Growth plan by end of the weekend and you’ll be moved off the waitlist next week (if you’re already on Growth/Team/Enterprise etc then you’ll be moved off waitlist too)

I missed it unfortunately! Looking forward to watching the recording please!

crazy

Buildprint is brilliant. Great job @georgecollier . Upgrading to Pro shortly

@georgecollier Could you add the ability to give the token a label so we know why it was created?

Buildprint live deployment opt-in

Buildprint can now deploy to live itself. This is an opt-in feature available in Project settings.

Users are opted out by default.

Feature request: I have two ChatGPT/Codex subscriptions. I would love to connect both, and just toggle once I hit my limit.

For now just disconnect and connect a new one

But $200/mo plan (even $100/mo) on Codex should be more than enough for all but the most active power users?

Automated testing revamped

Big update time!

Tests in Buildprint were cool but kind of slow, expensive, and unreliable. All bad things for automated tests.

Testing has just been rebuilt from the ground up. For now, you can still view and run legacy tests in the UI, but will not be able to build new ones.

Tests as code

  • Tests are now defined as code
  • They live alongside your Bubble project so agents can see and update them when they make changes
  • Tests are programatic rather than agent driven, making them fast, token efficient, and reliable
  • Tests can optionally have an agent step which lets the agent make judgment about a test’s success criteria

Run tests via the CLI

For now, we’ve removed the ability to automatically run tests from Buildprint Web. Instead, you ask the agent to run tests for you. This lets an agent easily test its work after making changes.

View tests and runs in web

Whilst tests run via CLI, you can monitor and debug them via web. Each test includes screenshots and a video recording of the end to end test.

Best practices are available here.