Show HN: Dropbase AI – A Prompt-Based Python Web App Builder(github.com)
github.com
Show HN: Dropbase AI – A Prompt-Based Python Web App Builder
https://github.com/DropbaseHQ/dropbase
3 comments
Ooooh, this is interesting. I think I'm building something quite similar! May I ask, how do you solve the code modification problem? In your demo video it shows the AI prompt is modifying code, not just generating it first-time, but I am curious how you do it. Are you using diffs?
I wrote about my approach here using ASTs: https://www.codeplusequalsai.com/static/blog/prompting_llms_...
You wrote in your post that you 'regenerate' a file - is that how you do it? Is it reliable? How does that work on big files? Does it fail at reproducing the rest of the file that should remain unchanged sometimes?
Thanks for answering any of these! Great project!
I wrote about my approach here using ASTs: https://www.codeplusequalsai.com/static/blog/prompting_llms_...
You wrote in your post that you 'regenerate' a file - is that how you do it? Is it reliable? How does that work on big files? Does it fail at reproducing the rest of the file that should remain unchanged sometimes?
Thanks for answering any of these! Great project!
Thank you! I read your blog post and checked out your project! If I understood it correctly, you’re trying to build a software engineering team in a box. Basically from first issue, to code, to live apps. Very interesting approach adding the collaboration angle! ASTs are neat but I’d imagine it could get hard to manage with more complex code.
In our case, we regenerate the `main.py` file each time. One of the hacks we did was to start with boilerplate code, which is why you see it modifying the code as opposed to generating from scratch the first time. We also feed the model with some context/rules on app building using our web framework, so the output is more bounded.
We haven’t tested it on really big files yet, though I'd imagine it could be a problem later. At the moment, we don’t generate HTML, JS/TS, or React code from scratch so our files tend to be relatively smaller than if we did. Our UI is defined via the `properties.json` file, which abstracts much of the underlying code, therefore keeping the files small. It’s much easier for LLMs to generate json and map it to UI behavior, than generate of the client code needed to do all of it.
We don’t have issues with the LLMs changing function/method code, but it occasionally implements one of boilerplate methods we didn’t explicitly ask for. In those cases, a developer has to remove that code manually, which is why showing code diff is critical.
Many other hacks come down to lots of prompt engineering! Something along the lines of "Only implement or modify a method/function corresponding to a user's prompt. Leave all others intact"
Happy to chat more!
Also you might find this blog post we wrote interesting: https://www.dropbase.io/post/an-internal-tools-builder-that-...
In our case, we regenerate the `main.py` file each time. One of the hacks we did was to start with boilerplate code, which is why you see it modifying the code as opposed to generating from scratch the first time. We also feed the model with some context/rules on app building using our web framework, so the output is more bounded.
We haven’t tested it on really big files yet, though I'd imagine it could be a problem later. At the moment, we don’t generate HTML, JS/TS, or React code from scratch so our files tend to be relatively smaller than if we did. Our UI is defined via the `properties.json` file, which abstracts much of the underlying code, therefore keeping the files small. It’s much easier for LLMs to generate json and map it to UI behavior, than generate of the client code needed to do all of it.
We don’t have issues with the LLMs changing function/method code, but it occasionally implements one of boilerplate methods we didn’t explicitly ask for. In those cases, a developer has to remove that code manually, which is why showing code diff is critical.
Many other hacks come down to lots of prompt engineering! Something along the lines of "Only implement or modify a method/function corresponding to a user's prompt. Leave all others intact"
Happy to chat more!
Also you might find this blog post we wrote interesting: https://www.dropbase.io/post/an-internal-tools-builder-that-...
It seems to me the more killer product here is the "Writing two files to build a webapp", and you could comfortably rip out ChatGPT and market to a wider audience?
I like your take on that! I hadn't thought about it that way before but "Writing two files to build a webapp" indeed sounds quite intriguing. And we could extend that idea to "...and deploy it with 1 click" or some version of that.
I'm curious about what audience you have in mind and what kind of apps would you be interested in building this way? Would love to hear more of your thoughts!
Edit: I should add that our main motivation for integrating GPT is that we had to introduce some new concepts to make this experience work, which increased the app-building learning curve. We thought having GPT generate code and highlighting diffs would be a neat way to teach users how to develop apps without reading a lot of documentation.
I'm curious about what audience you have in mind and what kind of apps would you be interested in building this way? Would love to hear more of your thoughts!
Edit: I should add that our main motivation for integrating GPT is that we had to introduce some new concepts to make this experience work, which increased the app-building learning curve. We thought having GPT generate code and highlighting diffs would be a neat way to teach users how to develop apps without reading a lot of documentation.
Aha, thanks for that detailed answer! Really fascinating to hear others' approaches to this area of building simple but full apps with LLMs. I'll definitely be following your progress, curious to see where this goes. And I will read your blog post this afternoon!
hi cryptoz! I'm curious to read about your approach but it seems I'm not the only one, your website went offline.
Hm, might be a DNS issue, not sure. I'll look into it, thanks!
Loaded fine for me who just clicked it a few minutes ago
Can it generate a django app? Or a Litestar app? Or any kind of framework?
No. At the moment all apps work on top of our internal framework. However, you can reuse your existing Python code/libraries. For instance, you could call existing Django models from Dropbase apps. You can also import any PyPI package.
This looks fantastic! Excited to try Dropbase for building quick, customizable web apps with ease.
Dropbase is an AI-based Python web app builder.
To build this, we had to make significant changes from our original launch: https://news.ycombinator.com/item?id=38534920. Now, any web app can be entirely defined using just two files: `properties.json` for the UI and `main.py` for the backend logic, which makes it significantly easier for GPT to work with.
In the latest version, developers can use natural language prompts to build apps. But instead of generating a black-box app or promising an AI software engineer we just generate simple Python code that is easily interpreted by our internal web framework. This allows developers to:
(1) See and understand the generated app code. We regenerate the `main.py` file and highlight changes in a diff viewer, allowing developers to understand what exactly was changed.
(2) Edit the app code: Developers can correct any errors, occasional hallucinations, or edit code to handle specific use cases. Once they like the code, they can commit changes and immediately preview the app.
Incidentally, if you’ve tried Anthropic’s Artifacts to create “apps”, our experience will feel familiar. Dropbase AI is like Claude Artifacts, but for fully functional apps: you can connect to your database, make external API calls, and deploy to servers.
Our goal is to create a universal, prompt-based app builder that’s highly customizable. Code should always be accessible and developers should be in control. We believe most apps will be built or prototyped this way, and we're taking the first steps towards that goal.
A fun fact is that model improvements were critical here: we could not achieve the consistent results we needed with any LLM prior to GPT-4o and Claude 3.5 Sonnet. In the future, we’ll allow users to modify the code to call their local GPT/LLM deployment via Ollama, rather than relying on OpenAI or Anthropic calls.
If you’re building admin panels, database editors, back-office tools, billing/customer dashboards, and internal dev tools that can fetch data and trigger actions across any database, internal/external service or API, please give Dropbase a shot!
We're excited to get your thoughts and questions!
Demos:
- Here’s a demo video: https://youtu.be/RaxHOjhy3hY
- We also introduced Charts (beta) in this version based on suggestions from cjohnson318 in our previous HN post: https://youtu.be/YWtdD7THTxE
Useful links:
- Repo here: https://github.com/DropbaseHQ/dropbase. To setup locally, follow the quickstart guide in our docs
- Docs: https://docs.dropbase.io
- Homepage: https://dropbase.io