Hey everyone,
I'm working on a new approach to AI-assisted coding that uses a visual, graph-based interface (similar to Unreal Engine's Blueprint system) instead of traditional text prompts.
TLDR: I am looking to see if my idea will be useful to anyone, I would greatly appreciate any responses.
The Problem
Current AI coding tools require managing context through text, making it difficult to build larger projects and visualize dependencies. I've found myself struggling to maintain context across multiple interactions and spending too much time crafting perfect prompts.
The Solution
Features:
- A split-screen showing both the visual graph and generated code
- Three main node types (class, function, variable) that you connect visually
- Project-specific AI agents that maintain context
- Visual dependency mapping between nodes
- You describe node functionality in natural language; AI generates the code
Additional Features
- Visual debugging with runtime value inspection
- Version control with branch visualization
- Automated documentation from node descriptions
- Code import/export system
I'd Love Your Feedback
Would this approach solve any problems in your current development workflow? What features would make this most valuable to you?
Thanks for your time!
[link] [comments]