Designing with Agentic AI in Blender

Three connected courses on designing with AI agents that drive Blender through MCP. Not Blender tutorials — training in how to prompt agents, catch their failure modes, and keep human judgment in the loop.

The same artifact — say, an office chair — travels all three courses: designed, then optimized for a game, then rigged to move. Each course runs beginner → master.

The three courses

Course 1

Product Design with Agentic AI

6 modules

High-quality static models — chairs, lamps, appliances — designed by prompting AI agents in Blender.

Learn to specify intent precisely enough that an agent builds a clean, correctly-scaled model, and to catch the constraints it silently drops.

↓ same artifact continues to Game Asset

Course 2

Game Asset Optimization with Agentic AI

4 modules

Take the same models and make them game-ready: low-poly retopology, UV unwrapping, and baked maps.

The optimization stage is where agents are strong but quietly lossy. Learn to drive budgets, UVs, and bakes — and to measure what was lost.

↓ same artifact continues to Rigging & Animation

Course 3

Rigging & Animation with Agentic AI

4 modules

Make the same models move — and discover where the agent capability boundary actually is.

Rigging is the hardest stage for agents. Learn to push them to their limit, recognize where they break, and take over with judgment.

Every course runs the same arc

Beginner

First prompts, basic constraints, reading agent output.

→

Intermediate

Multi-step prompts, references, recovering from common failures.

→

Advanced

Repeatable workflows, quality gates, edge cases the agent fumbles.

→

Master

The capability boundary: where you stop trusting the agent and take over.