BlenderRAG: High-Fidelity 3D Object Generation via Retrieval-Augmented Code Synthesis

TL;DR

BlenderRAG enhances the automatic generation of Blender code from natural language by using retrieval-augmented techniques on a curated dataset, significantly improving success rates and semantic alignment without fine-tuning.

Contribution

It introduces a retrieval-augmented system that leverages a curated multimodal dataset to improve code generation accuracy and semantic consistency in 3D object creation.

Findings

Compilation success rate increased from 40.8% to 70.0%.

Semantic alignment improved from 0.41 to 0.77 (CLIP similarity).

System works across four state-of-the-art LLMs without fine-tuning.

Abstract

Automatic generation of executable Blender code from natural language remains challenging, with state-of-the-art LLMs producing frequent syntactic errors and geometrically inconsistent objects. We present BlenderRAG, a retrieval-augmented generation system that operates on a curated multimodal dataset of 500 expert-validated examples (text, code, image) across 50 object categories. By retrieving semantically similar examples during generation, BlenderRAG improves compilation success rates from 40.8% to 70.0% and semantic normalized alignment from 0.41 to 0.77 (CLIP similarity) across four state-of-the-art LLMs, without requiring fine-tuning or specialized hardware, making it immediately accessible for deployment. The dataset and code will be available at https://github.com/MaxRondelli/BlenderRAG.