Art Tutor Fundamentals

Practice Zone

Inspiration

Many aspiring artists hit a "wall" when transitioning from 2D sketches to 3D depth. Understanding perspective often involves dry theory or static diagrams that don't translate easily to the page. We wanted to build a bridge between digital 3D space and traditional drawing by giving artists a tool for spatial reasoning that eventually becomes second nature.

What it does

Art Fundamentals Tutor is an interactive training ground for mastering perspective.

The Sandbox: Users manipulate a real-time 3D workspace to visualize 1, 2, and 3-point perspective.
The Challenge: Users draw their own version of the 3D scene on paper or digital canvas.
The Critique: Using Gemini Vision, the app analyzes a photo of the user's drawing, compares it against the 3D wireframe, and provides an instant grade and constructive feedback.
The Habit: A gamified leveling system with XP, streaks, and a personal gallery keeps artists motivated to practice daily.

How we built it

Development Platform: Built entirely within Google Antigravity, leveraging its agentic development workflows to orchestrate complex tasks across the editor and browser.
Frontend & 3D: Developed with React/Next.js and Three.js to create the interactive 3D workspace, allowing for real-time manipulation of geometry and vanishing points.
AI Integration: Integrated Google Gemini for high-accuracy visual analysis. We utilized Antigravity's agent-first environment to streamline the integration of Gemini Vision, ensuring the AI could accurately interpret 3D metadata alongside user uploads.

Challenges we ran into

Aligning the "Real" and "Virtual": Teaching the AI to accurately compare a hand-drawn 2D image to a perfect mathematical 3D model.

Accomplishments that we're proud of

The Wireframe Overlay: Developing a system where the AI can visually point out exactly where a user’s lines went "off-grid" by overlaying the 3D wireframe onto their uploaded sketch.
Agentic Efficiency: Successfully using the Antigravity IDE to reduce context switching, allowing us to build the full-stack application and AI critique zone.

What we learned

AI as a Partner, Not a Replacer: We learned that AI is most effective in art education when it provides specific, technical feedback rather than just subjective praise.
The Power of Agentic Workflows: Developing in an agent-first era with Google Antigravity taught us how to delegate repetitive verification tests to AI agents, letting us focus on the creative UX.

What's next for Art Tutor Fundamentals

Complex Geometry: Adding primitives like cylinders and spheres for more advanced perspective training.
Anatomy Module: Using the same 3D-to-AI feedback loop to teach human proportions.