Awesome video. It's interesting, informative, and entertaining.
Founders I talk to that are doing hardware, broadly speaking, say it's a competitive advantage as it's not as crowded. Content like Aaed's will hopefully nudge more people into it.
The breach of trust here, which is hard to imagine isn't intentional, is enough reason alone to stop using Vercel, and uninstall the plugin. That part is easy. Most of these agents can help you migrate if anything.
Hi HN, I’m one of the founders of FlutterFlow (YC W21). We launched Dreamflow earlier this year.
What it is: Dreamflow generates a Flutter app from a prompt, then lets you keep working in a real dev environment with a visual editor, AI agent, and full code (code is the source of truth).
Why we built it:
1) make small changes dead simple: tweak properties, move widgets, add components without editing code or prompting; and
2) give the agent precise context: select a widget to add it to the prompt (screenshot included automatically); the agent has access to read runner logs.
What’s different:
- Bi-directional editing: canvas ↔ code; standard Flutter.
- Runs in the browser: hot reload, preview.
- Agentic edits with diffs; widget-aware context + log access.
How to try:
Prompt: “Fundraising CRM for founders: pipeline (Targeted -> Closed); investors (name/firm/check size/intro source); notes + next steps + due date; goal $X/$Y; tabs Pipeline/Investors/Tasks/Settings”
Then preview, rearrange visually, integrate a backend or keep it local, and export/publish.
Current gaps:
Visual coverage isn’t complete; GitHub integration is in progress; importing existing Flutter projects isn’t supported yet.
We’d love your feedback. Where does the visual editor help or get in your way? Which integrations would you expect first?
Agreed, making it more natural to create apps in a way that follows Flutter best practices is a desirable end goal. There's so much to improve, it's exciting :-)
It's not quite an apples-to-apples comparison, but the memory use of FlutterFlow is comparable to WebFlow. Our landing page is built using WebFlow, and I compared memory use of that to a comparably sized project in FF.
Performance could certainly be improved (especially for Web), but it's gotten much better over the last year.
Thanks for the suggestions! Yes, there's definitely room for improvement with the generated code ... I think the padding -> SizedBox conversion at least in some cases (only when main-axis padding is added) can be automated. We can experiment with code improvement features such as that and see how it works.
Expanded + Percentages for widths are possible already
There were a lot of issues with custom widgets when we first launched the feature, we've deployed several fixes since then. But there's still more to do on our end to make it clear what errors you're getting. Thanks for the suggestion!
The reasons we prioritized Firebase is because it works really well with Flutter and was also what most of our early users preferred. FlutterFlow itself is in fact a Flutter+Firebase app. We take the best ways we know how to build Flutter+Firebase apps and make it easy to do, in a browser.
We definitely want to support more DBs in the future, just a matter of time. Meanwhile, we see some users achieve that goal using API integrations and code export
Yep, you can push the latest changes directly to Github. We always use a specific branch "flutterflow". Users can then merge in changes, as long as the changes on the main branch are not changing the layout visuals significantly, Git handles the merge nicely.
Ofc you can't import code into FF from your main branch, at least at the moment. But it's still been working well for our users
The reason there are 3 different kinds of custom code is so that it can easily integrate with the visual builder ... for instance a "Custom Action" function will show up when you add an action on a Button Tap, or TextField submit. A "Custom Widget" can be dragged in and passed parameters, and a "Custom Function" can be used to set values. Would love any feedback
Thanks! We're gonna work hard to get it to a place where the set up you love and are used to is also available in FF (or vice versa), and that you get the best of both worlds.
Thanks for your feedback! The reason we don't have this yet is only because we're a small team, but we're growing and will definitely have NGO/non-profit discounts!
Re: AppGyver, one big difference is the ability to generate clean code that can be exported. And there's also Flutter, which we think is the framework that will finally make low-code app builders work.
Being able to easily search and find where your network requests, custom function calls, conditional variables, etc are used is definitely something that has come up. We've started making progress by adding visual elements that help you identify them. For instance, you can see what element on the widget tree has a query (or action) on it based on an icon + color. We plan on adding toggles for overlays of different views, along with the ability to see where an API call is used (from the API configuration page), or where a specific collection is read (from the Firestore configuration page). It's definitely solvable, and we'll keep talking to users and iterating till we get there :)
thanks! we have some exciting features coming up.
- deploy to app stores from FlutterFlow, via Codemagic
- search functionality, via Algolia
- several widgets (map/chat/pop-ups/bottom sheet)
in terms of our team, we're looking to bring on a community manager and a content strategist.
Founders I talk to that are doing hardware, broadly speaking, say it's a competitive advantage as it's not as crowded. Content like Aaed's will hopefully nudge more people into it.