obra

Scan To Model — AI skill for Claude Code

AI community

Evidence-driven scan reconstruction skills and local tools for Codex and Blender.

How to install Scan To Model

This entry records only its repository, not the path inside it, so there is no exact command to give. Open obra/scan-to-model and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What Scan To Model does

Evidence-driven scan reconstruction skills and local tools for Codex and Blender.

Alternatives in AI

  • Codeburn — Free, local tool to track AI coding token usage and cost across 37 tools and agents (Claude Code, Cursor, Code 9.7k ★
  • Academic Humanizer — Strip AI-writing tells from papers and grant proposals (NSF/NIH), while keeping scholarly voice and tying clai 1.1k ★
  • Meshgen — Use AI Agents directly in Blender 914 ★

README

Scan to Model

A Codex plugin for turning scan evidence into reviewable Blender architectural models.

Start with photographs, calibrated depth, camera poses, plans, and observations. Preserve the evidence, measure real surfaces, align captures, and build a model whose geometry can be traced back to its sources. The plugin provides seven skills and local Python tools. It does not automatically reconstruct an entire building from a ZIP.

Install in Codex

Add this repository as a marketplace and install its plugin:

codex plugin marketplace add obra/scan-to-model
codex plugin add scan-to-model@scan-to-model

Start a new Codex task so it can load the installed skills, then ask:

Use scan-to-model to inventory these scans, build a coverage atlas, and model the next evidence-supported room, including its ceiling and connections.

The workflow runs locally. It needs no server or API key. Supply your own project directory and scans; keep private evidence in that project.

What it does

Skill Result
scan-to-model An evidence hierarchy, a coverage atlas, and a bounded modeling plan.
scan-ingest Preserved source archives, pairing and variant audits, frame indices, and contact sheets.
source-observe Annotate visible construction in original photographs, retaining uncertain identity and endpoints independently of model placement.
street-view-reference Verified exterior panorama references with imagery dates, viewpoint, attribution and explicit limits on dimensional use.
depth-inspect Calibrated native point references and measurements tied to photographed physical surfaces.
scan-register Unit-scale rigid alignment with withheld checks and explicit limits on where a transform is valid.
blender-reconstruct Source-tagged model candidates, saved-scene inventories, comparison renders, and a review before promotion.

The guidance keeps current architecture, historical concealed constructi