forgecad-image-replicator

Build ForgeCAD geometry from reference images with multi-view validation.

917|102|Updated Feb 8, 2026

Search Tags:#validation #parametric modeling #3d reconstruction #image to cad #render comparison #ForgeCAD #evidence-based inference

Ready to run

npx skills add https://github.com/KoStard/forgecad-public-kit --skill forgecad-image-replicator

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: forgecad-image-replicator
Source: https://github.com/KoStard/forgecad-public-kit/tree/main/skills/forgecad-image-replicator
Command: npx skills add https://github.com/KoStard/forgecad-public-kit --skill forgecad-image-replicator

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires puppeteer-core, and includes scripts (resource) components.

What problem does it solve?

ForgeCAD Image Replicator solves the problem of converting one or more photos into a believable, parametric 3D CAD model without only matching a single camera angle.

Core Features & Use Cases

3D object first, image match second: Builds a Real Object Brief (identity, scale, manufacturing posture, hidden-side expectations) before optimizing camera matching.
Evidence-driven inference across multiple views: Treats each provided image as evidence and validates against all usable reference cameras, not just the “best” image.
Model validation with canonical and reference-camera renders: Produces side-by-side comparison boards and includes inspection-oriented checks for multi-part and fit-sensitive geometry.

Quick Start

Use forgecad-image-replicator to infer and build a ForgeCAD model from the attached reference images.

Frequently Asked Questions about forgecad-image-replicator

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I convert a reference photo into a parametric 3D CAD model?▼

Converting a photo to 3D CAD requires inferring underlying geometry from reference images and validating it against multiple camera views. This ensures scale and multi-view consistency for a believable parametric model.

What is the best way to ensure hidden surfaces are accurate during image to CAD reconstruction?▼

Accurate 3D reconstruction of hidden surfaces requires building a Real Object Brief first. This establishes identity, scale, and manufacturing posture before optimizing camera matching, ensuring hidden-side expectations are met.

How does multi-view validation work for photo-to-CAD workflows?▼

Multi-view validation treats each provided image as evidence and validates the inferred 3D geometry against all usable reference cameras. It produces side-by-side rendered comparison boards rather than matching a single camera angle.

Do I need camera calibration for every reference image when building 3D geometry?▼

Yes, explicit camera calibration per usable reference image is required. This deterministic blockout-to-detail iteration ensures the inferred ForgeCAD geometry aligns properly with each perspective perspective view.

Can I use ForgeCAD for evidence-based inference with only a single photo?▼

While a single photo can start the process, evidence-based inference across multiple views is necessary to satisfy scale and multi-view consistency. The workflow uses inspection-backed acceptance criteria to validate the final model.

What are the limitations of matching only a single camera angle for 3D reconstruction?▼

Matching a single camera angle often produces inconsistent 3D geometry with incorrect hidden surfaces. Applying reference-matched and canonical views with inspection-oriented checks resolves multi-part and fit-sensitive geometry issues.