Voice → Model → Print

Turn spoken ideas into printable objects.

This prototype orchestrates real-time speech, LLM intent extraction, cloud 3D generation, mesh repair, and slicing — all without local GPU compute.

PipelineServerless
3D GenerationMeshy (default)
OutputGLB + G-code

Live Voice Session

Speak your design intent

Push to talk and describe the object you want to print. The system will extract a prompt, generate a 3D model, repair it, and slice it for G-code.

Browser STT not supported here.
Examples: “box 80x40x20 mm, hollow, wall 3 mm”, “cylinder dia 40 mm height 60 mm, hole 10 mm”, “sphere 50 mm, rounded 2 mm”, “phone stand angle 65 deg”.
Uses the same pipeline as voice.
Pipeline Status
idle
Prompt will appear here.
Live Transcript
Waiting for input...

3D Preview

Inspect the generated model

Rotate, zoom, and validate the geometry before downloading the print-ready G-code.

No model yet
Speak a design intent to generate a printable model.
Download G-codeG-code link appears after slicing.