CodeVibe

When AI Falls Short, Humans Step In.

Comment

Inspiration

While AI code editors are transforming the way we build software, they’re not perfect — especially for beginners or non-tech users. Many face security loopholes, unclear errors, or AI limitations. We realized there's a gap: AI coders need real-time human support without breaking the "AI workflow vibe." That’s where Code Vibe was born — to bridge AI efficiency with human expertise.

What it does

Code Vibe is a collaborative platform for AI-powered coders (we call them vibe coders) to connect with real developers when they hit a roadblock.

Here’s how it works:

Vibe coders post their code problems, snippets, and language/tool used.

Developers with matching skills are notified and can bid on the problem — stating how they'll solve it and their estimated time.

Vibe coders can accept or reject bids.

Once accepted, both parties enter a secure chat window (no phone numbers or links allowed).

If needed, they can generate a Google Meet link to solve the problem live.

In future, an AI bot will monitor calls to ensure compliance and prevent off-platform communication.

How we built it

Frontend: React.js (for responsive and interactive UI)

Backend: Node.js with Express (for REST APIs and real-time updates)

Database: Supabase (for user auth, problem-posting, bid management)

Real-time functionality: Supabase Realtime and WebSockets

Chat system: Custom-built with filtering logic to block sensitive info

Video meet integration: Google Meet API

Deployment: Netlify

Challenges we ran into

Real-time notifications: Ensuring that only relevant developers get notified without spamming others.

Message filtering: Designing a system to block phone numbers, links, and personal info within the chat.

Bid logic: Creating a clear bidding system that feels natural and fair for both coders and developers.

Security design: Preventing misuse or off-platform connections, especially before our future AI bot is ready.

Accomplishments that we're proud of

Built a complete end-to-end working prototype during the hackathon.

Designed a real-time developer bidding system — something rare in AI-code collab platforms.

Integrated chat and video seamlessly while maintaining platform safety.

Created a new kind of coder-developer connection model centered around AI workflows.

What we learned

How to design scalable real-time systems with Supabase.

Importance of building platforms that combine AI support with human touch.

How human interactions in coding platforms need structure, moderation, and fallback options like meet links.

Balancing user freedom with safety when designing open platforms.

What's next for CodeVibe

Integrate our AI bot that will join video calls and detect violations like sharing personal info or external contact.

Add payment integration so developers can offer premium support.

Create a reputation system for both vibe coders and developers based on ratings and completed tasks.

Expand to support open-source contributions and project-based collaboration.

Launch a mobile app version to help coders on the go.

Built With

Share this project:

Updates