A Claude Code skill that runs a scripted pipeline to produce product films and packshots from a design patent or photos, driving Blender, Octane and DaVinci Resolve.
yunfanye/openhouse-3d
Python toolkit that rebuilds a house from real-estate listing photos as a 3D Blender model and renders a walkthrough video.
Open-source alternative to:Matterport
- Stars
- 60
- Forks
- 8
- Open issues
- 0
- Last commit
- last month
What is it?
OpenHouse 3D is a modeling toolkit for turning the photos in a real-estate listing into a procedural 3D house in Blender (bpy). It solves each listing photo's camera from image features, including perspective-corrected shots and drone aerials, and builds walls with validated openings, siding, brick, windows, doors, stairs, roofs with gutters, and fixtures. Furniture, rugs, vegetation, and neighbouring context can be added for virtual staging.
The project renders with headless Cycles on CPU, Metal, OptiX, CUDA, HIP, or oneAPI. It produces a continuous cinematic walkthrough at 1080p24 with title cards and music, a gallery of photo-matched renders, and a synchronized video that plays the film beside the original photos. Scenes can be packed, rendering can resume in chunks, and camera routes are audited for clearance.
The project is a toolkit rather than a one-click service: a person or AI coding agent interprets the photos and writes a house package, and the library handles geometry, rendering, and video. Hidden geometry is inferred, so the output is a reconstruction rather than a measured survey. Example houses and a synthetic starter template are included.
Key features
- 01Solves camera poses for listing photos, including shifted-lens shots and drone aerials.
- 02Procedurally builds walls with validated openings, siding, brick, windows, doors, stairs, and roofs.
- 03Adds furniture, rugs, art, vegetation, and surrounding context for virtual staging.
- 04Renders headless Cycles on CPU, Metal, OptiX, CUDA, HIP, or oneAPI with resumable chunked output.
- 05Produces a 1080p walkthrough video, a photo-vs-render gallery, and a synchronized comparison movie.
Use cases
- Creating a virtual walkthrough of a property from its existing listing photos.
- Producing photo-matched renders to compare a 3D model against the real house.
- Prototyping architectural visualizations in Blender with scripted, reproducible scenes.
- Studying how camera solving and procedural modeling can be combined in a Blender pipeline.
Quick start
- 1
Clone the repository
git clone https://github.com/yunfanye/openhouse-3d.git - 2
Install the Python dependencies in a virtual environment
python -m pip install -e '.[dev]' - 3
Create the synthetic starter house
python tools/new_house.py my_house --title "My house" - 4
Render a still image of the starter house with headless Blender
blender -b --python-exit-code 1 --python run.py -- --house my_house --cam hero --samples 16 --scale 25 --no-polish --device CPU
Things to know
- Requires Blender 5.2.x and a local Python 3.9+ setup.
- It is a modeling toolkit, not a one-click photo-to-3D service, and needs a person or AI agent to build each house package.
- Reconstructions are inferred from photos and are not measured surveys.
In collections
Topics
FAQ
- What is openhouse-3d?
- Python toolkit that rebuilds a house from real-estate listing photos as a 3D Blender model and renders a walkthrough video. OpenHouse 3D is a modeling toolkit for turning the photos in a real-estate listing into a procedural 3D house in Blender (bpy). It solves each listing photo's camera from image features, including perspective-corrected shots and drone aerials, and builds walls with validated openings, siding, brick, windows, doors, stairs, roofs with gutters, and fixtures. Furniture, rugs, vegetation, and neighbouring context can be added for virtual staging.
- Is openhouse-3d free to use? What license does it have?
- openhouse-3d is open source under the MIT license. Check the license terms before using it in commercial products.
- Is openhouse-3d still maintained?
- Updated within 6 months. The last commit to openhouse-3d was on Sep 26, 2026.
- What can I use openhouse-3d for?
- Creating a virtual walkthrough of a property from its existing listing photos. Producing photo-matched renders to compare a 3D model against the real house. Prototyping architectural visualizations in Blender with scripted, reproducible scenes. Studying how camera solving and procedural modeling can be combined in a Blender pipeline.
- How do I install openhouse-3d?
- Clone the repository: git clone https://github.com/yunfanye/openhouse-3d.git Install the Python dependencies in a virtual environment: python -m pip install -e '.[dev]' Create the synthetic starter house: python tools/new_house.py my_house --title "My house" Render a still image of the starter house with headless Blender: blender -b --python-exit-code 1 --python run.py -- --house my_house --cam hero --samples 16 --scale 25 --no-polish --device CPU
- Where can I get openhouse-3d?
- The source code and installation instructions are on GitHub: https://github.com/yunfanye/openhouse-3d
Summary generated by AI from the project's README. GitHub stats refresh daily.
Related projects
An agent-driven skill that turns one interior room photo into an editable Blender scene, a Three.js viewer and a portable GLB model.
A general-purpose Structure-from-Motion and Multi-View Stereo pipeline with GUI and command-line interfaces for 3D reconstruction from images.
Open-source node-based visual programming tool for photogrammetry, turning photos into 3D models, camera tracking and HDR images.
An iPhone LiDAR capture app plus Three.js viewer and agent skill for rebuilding rooms as editable 3D scenes.
Cross-vendor 3D Gaussian Splatting trainer that turns photos or video into splats and textured meshes, running on Vulkan or CUDA without Python.