VibeDB

Inspiration

When we first started working on full stack projects with databases we had absolutely no knowledge of SQL. Documentation was unclear, services were confusing to set up, and even supposed beginner friendly platforms didn't provide a level of abstraction needed for us to set up our queries. We know we weren't the only ones who struggled with this, as there are many developers out there who aren't technical or just are looking for faster ways to set up their backend. These experiences inspired our team to create VibeDB, an app that fills in this gap by allowing users to seamlessly create complete databases without writing a single line of SQL code.

What it does

VibeDB is able to convert a project description into a schema visualization with rows and columns inferred by prompts with the help of the OpenAI api. With a button on the corner of the screen, you can easily extract the SQL query code for your schema into your clipboard, or use the visual editor to make quick revisions before pasting it to any database platform i.e. Supabase, pgadmin, etc.

How we built it

We utilized the OpenAI api for generating the SQL schemas from user prompts. For data formatting we used response URLs from a FastAPI server. For the front end pipeline we utilized NextJS.

Challenges we ran into

Creating such a profound app is a tough process and one major obstacle we had to overcome early on was designing the app in a straightforward way that we can display tables containing any type of object extracted from random OpenAI response strings. Technical issues involved transferring AI responses from one page to another page, fixing missing characters left out of backend responses, and making sure our table interfaces could work.

Accomplishments that we’re proud of

Throughout the development process, we embraced the dynamic nature of building something innovative and successfully navigated various challenges that pushed us to think creatively and adapt our approach. We're incredibly proud of what we accomplished with VibeDB and excited about how it meaningfully addresses a real need in the developer community that previously went unmet.

What we learned

We learned how to integrate FastAPI with ChatGPT's API, handle complex React state for drag-and-drop interfaces, and build a tool that actually solves developer workflow problems.

What’s next for VibeDB

We're thrilled about the roadmap ahead for VibeDB, shaped by constructive feedback from judges and mentors. We're looking forward to adding the ability to link the app with an MCP, potentially for Supabase integration. We're also excited to develop a more robust editor that will allow users to submit multiple prompts, enabling them to iteratively build upon their existing database structures and create even more sophisticated solutions.

Built With

• css
• fastapi
• nextjs
• openai
• python
• typescript