gaussian-splat-to-blender-nodes — Claude Code plugin
gaussian-splat-to-blender-nodes is a Claude Code plugin that guides Blender scene reconstruction from a Gaussian splat using editable geometry and shader nodes. It can also map calibrated capture textures onto the result.
MIT ★ 86 updated 2026-10-24
/plugin marketplace add gokayfem/gaussian-splat-to-blender-nodes
/plugin install gaussian-splat-to-blender-nodes@gaussian-splat-to-blender-nodesWhat gaussian-splat-to-blender-nodes does
The plugin guides an agent through measuring a source scene, dividing it into parts, and building those parts with Blender Geometry Nodes and Shader Nodes. This is a reconstruction workflow, not a one-click splat converter. The resulting objects use editable geometry, parameters, and materials.
The toolkit includes asset inspection for PLY, SPLAT, and KSPLAT metadata, camera utilities, scene audits, and render comparisons. Appearance mapping tools handle UV preparation, depth and object visibility, projection, and packed hybrid-scene assembly. A scan-free starter run requires Python 3.10+ and Blender 5.0+; Blender versions beyond the tested 5.0.1 need separate validation. Structural audit checks do not establish visual resemblance.
What it helps with
- rebuild editable Blender geometry from a Gaussian splat
- break a scene into parts and assemble them with node graphs
- create repeatable camera views for a reconstruction
- compare rendered views against captured references
- map captured appearance onto reconstructed objects
Example prompts for the agent
- Rebuild this Gaussian splat as editable Blender geometry and review the scene from several views.
- Create Geometry Nodes and Shader Nodes for this scene, then compare the renders with the reference views.
- Apply calibrated capture-texture mapping to the reconstructed objects.
How to install gaussian-splat-to-blender-nodes
Plugin marketplace
/plugin marketplace add gokayfem/gaussian-splat-to-blender-nodes/plugin install gaussian-splat-to-blender-nodes@gaussian-splat-to-blender-nodesStep by step, with commands for each platform: Claude Code plugins: marketplaces, installing, and building your own
Questions about gaussian-splat-to-blender-nodes
How do I install gaussian-splat-to-blender-nodes in Claude Code?
Run these commands in order: /plugin marketplace add gokayfem/gaussian-splat-to-blender-nodes, then /plugin install gaussian-splat-to-blender-nodes@gaussian-splat-to-blender-nodes.
What does gaussian-splat-to-blender-nodes do?
It guides an agent through rebuilding a Blender scene from a Gaussian splat with editable geometry and materials. The toolkit also includes scene audits and view-comparison tools.
What is required for a scan-free test run?
The starter workflow requires Python 3.10+ and Blender 5.0+. Other Blender versions need separate validation.
Source text
SKILL.md · MIT
# World to World Mapping
**Gaussian splat to Blender nodes — reconstruct geometry, then optionally map captured appearance.**
A project-independent agent skill and toolkit for measured reconstruction with native
**Geometry Nodes, Shader Nodes and calibrated multi-view texture projection**. Built from
the Garden study, in the spirit of [video-to-blender](https://github.com/gokayfem/video-to-blender)
and [shader-skills](https://github.com/gokayfem/shader-skills).
<p align="center">
<img src="gallery/world-to-world-mapping.jpg" width="800" alt="World to World Mapping release: grey and textured Blender scene comparison">
</p>
[Watch the release video — grey/textured reveals and a matched 360° comparison](https://v3b.fal.media/files/b/0aafb407/k_VXzVOBiFnE1RwnpVGxF_world_to_world_mapping-gaussian_splat_to_blender_nodes.mp4)
· [Read the skill](skills/gaussian-splat-to-blender-nodes/SKILL.md)
· [Reusable prompts](prompts/README.md)
· [Hybrid mapping guide](skills/gaussian-splat-to-blender-nodes/references/hybrid-projection.md)
*The video shows the accepted Garden c26b hybrid experiment. Its mapped table and paving
are baked evaluated meshes with packed appearance textures; other reconstructed parts retain
node-based construction. The source scan and fitted hybrid assets are not bundled. See
[media provenance and limitations](gallery/README.md). The installation name remains
`gaussian-splat-to-blender-nodes`.*
## What it does
This is an **agent-guided reconstruction workflow**, not a one-click splat-to-mesh converter.
The splat provides observations; the output describes objects through meaningful geometry,
parameters and materials that you can edit in Blender.
```text
Splat reference → measured frame + cameras → semantic scene parts
→ Geometry Nodes + Shader Nodes → native audit + visual comparison
→ optional calibrated appearance mapping → hybrid review + packed blend
```
| Included | Purpose |
|---|---|
| Installable agent skill | Acquisition, registration, construction, detail and candid critique |
| Project-agnostic prompt pack | Full brief, whole-scene capture, part contracts, detail, critique and delivery |
| Asset inspector | SHA-256 and limited PLY / SPLAT / KSPLAT metadata |
| Shared camera tools | Validated coordinates and repeatable world/scan orbit manifests |
| Native-scene audit | Remove the reference in memory; compare mesh and instance signatures |
| Blender renderer | Sequential rendering of a shared camera manifest |
| Still/video comparison | Strict frame pairing, labels and hashed evidence; optional FFmpeg MP4 |
| Synthetic starter | Small editable GN table; no scan download required |
| Hybrid mapping tools | UV preparation, depth/object visibility, robust projection, material-aware fallback and packed assembly |
| Hybrid prompt/config | Per-project objects, material slots, cameras, source masks and appearance mode |
| Garden study | Seven procedural parts, parameters and vendored node-building helpers |
No automatic semantic segmentation, universal splat decoder or reference renderer is bundled.
An audit pass proves specific structural checks; it does not prove visual resemblance.
## Install
### Agent Skills
```sh
npx skills add gokayfem/gaussian-splat-to-blender-nodes
```
### Claude Code plugin
```text
/plugin marketplace add gokayfem/gaussian-splat-to-blender-nodes
/plugin install gaussian-splat-to-blender-nodes@gaussian-splat-to-blender-nodes
```
### Manual / Codex
```sh
git clone https://github.com/gokayfem/gaussian-splat-to-blender-nodes.git
cd gaussian-splat-to-blender-nodes
mkdir -p ~/.codex/skills
cp -R skills/gaussian-splat-to-blender-nodes ~/.codex/skills/
```
For Claude project-local use, copy the skill into `.claude/skills/`. See
[adapters](adapters/README.md) for other agents. A skill-only install includes tools and the
starter; clone the full repository for Garden.
## Try it without a scan
Requires **Python 3.10+** and **Blender 5.0+**. Local integration verification used Blender
5.0.1; validate other versions before relying on them. Substitute your Blender executable
for `blender` if it is not on PATH.
```sh
S=skills/gaussian-splat-to-blender-nodes
blender -b --factory-startup --python-exit-code 1 \
--python "$S/templates/starter.py" -- --out outputs/starter.blend
blender -b --factory-startup outputs/starter.blend --python-exit-code 1 \
--python "$S/scripts/audit_blend.py" -- --out outputs/starter-audit.json
python3 "$S/scripts/cameras.py" --frames 4 --radius 3 --target 0 0 0.5 \
--width 640 --height 480 --out outputs/starter-cameras.json
blender -b --factory-startup outputs/starter.blend --python-exit-code 1 \
--python "$S/scripts/render_manifest.py" -- \
--manifest outputs/starter-cameras.json --out outputs/starter-views
```
Width and Depth controls change the tabletop and supports. The starter is a pipeline test,
not a reconstructed asset. The audit deletes its hidden synthetic reference point in memory
and leaves the saved file unchanged.
## Reconstruct your scene
Use the [full prompting guide](prompts/README.md), starting with the
[copy-ready project brief](skills/gaussian-splat-to-blender-nodes/templates/prompts/project-brief.md).
It generalizes the actual structure used for the study: source selection and whole-scene
capture, shared measurements/cameras, scoped part builders, integrated blockout, detail
correction and final evidence. It works for other scene types and with a single agent.
The [capture guide](skills/gaussian-splat-to-blender-nodes/references/whole-scene-capture.md)
explains how repeatable reference views came from the original splat: viewer/asset identity,
exact camera and canvas settings, settled captures, shared world coordinates, coverage views,
analysis maps and matched close-ups. This is reference capture, not splat training.
For a shorter starting message, give your agent an authorized source, intended views and a
quality target:
> Use gaussian-splat-to-blender-nodes. Reconstruct this scene as editable native Blender
> geometry and materials. First fingerprint the asset, establish a measured coordinate
> frame, match the reference camera and identify the major objects. Build an integrated
> coarse checkpoint before fine detail. Review side, opposite and detail views; be candid
> about mismatches. Deliver source, parameters, blend and scan-removal evidence.
Start with the [skill workflow](skills/gaussian-splat-to-blender-nodes/SKILL.md).
Use the [part contract](skills/gaussian-splat-to-blender-nodes/templates/prompts/part-contract.md),
[detail pass](skills/gaussian-splat-to-blender-nodes/templates/prompts/detail-pass.md) and
[review prompt](skills/gaussian-splat-to-blender-nodes/templates/prompts/review-and-correction.md)
as the project progresses.
For a turntable, follow the [matched 360° guide](skills/gaussian-splat-to-blender-nodes/references/comparison-videos.md).
Reference frames must come from a compatible viewer using the same camera manifest;
the Blender renderer handles the native side.
## Map captured appearance onto your reconstruction
Use the [hybrid mapping prompt](skills/gaussian-splat-to-blender-nodes/templates/prompts/hybrid-projection.md)
when captured appearance is part of the intended result. Supply your own reconstructed
scene and calibrated captures, then fill the
[configuration template](skills/gaussian-splat-to-blender-nodes/templates/projection.json).
Objects, materials, cameras, visibility masks and rejection rules are configuration inputs;
the implementation contains no required Garden objects or fitted scene coordinates.
The [step-by-step guide](skills/gaussian-splat-to-blender-nodes/references/hybrid-projection.md)
covers the three commands: **prepare → project → assemble**. Install `requirements.txt` for
the NumPy/SciPy/Pillow projector. Blender prepares evaluated UV surfaces and visibility maps;
ordinary Python maps the captures; Blender packs the results onto selected mesh copies.
Unseen/rejected texels retain the original material. Source-only occluders require masks or
other explicit evidence; reconstructed depth alone cannot identify them.
Choose unlit captured radiance or an approximate irradiance-divided lit material. Neither
recovers clean physical albedo. Selected mapped surfaces require rebaking after geometry or
UV edits; untouched procedural parts retain their live graphs. The original source is
preserved, and immutable manifests tie geometry, captures and outputs together.
## Garden example
Garden generates ground, table, vase/arrangement, tree, shrubs, boundary and structures
from nodes. It includes fitted parameters and a sparse foliage control field, not the
original Gaussian records. The default build requires no scan download.
<details>
<summary>View the table's editable Geometry Nodes graph</summary>

</details>
[Build instructions, provenance and limitations →](prompts/garden/README.md)
The reference is the Mip-NeRF 360 Garden scene displayed by GaussianSplats3D. Original scans,
decoded arrays and reference captures are not distributed here. No claim of perfect
replication is made: construction, foliage, surface response and light still differ visibly.
## Development
```sh
python3 -m pip install -r requirements.txt
python3 -m unittest discover -s tests -v
python3 tests/check_repository.py
```
Tests cover coordinate invariants, malformed assets, comparison pairing, projection visibility,
linear-light blending, material boundaries, fallback and artifact fingerprints. Blender checks
are documented in [tests](tests/README.md). Python CI alone does not validate rendering.
See [AGENTS.md](AGENTS.md) for contributor rules.
## Credits and license
MIT for repository code and instructions. Vendored nodekit/library retain their
[original license](prompts/garden/vendor/LICENSE.video-to-blender).
Datasets and third-party media retain their own terms; the code license does not grant
rights to those assets. See [CREDITS.md](CREDITS.md).
© gokayfem · MIT · gokayfem/gaussian-splat-to-blender-nodes