OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
Image & Video AI

Brickgpt

BrickGPT is a Python research project that generates physically stable, buildable toy-brick structures from text prompts.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Brickgpt — OpenViST project profile in Image & Video AI.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use BrickGPT when you want to turn a text description into a toy-brick structure and inspect the resulting model files. Choose local inference when you need the documented command-line workflow or want to adapt the Python project.

Who it is for

  • Researchers exploring text-guided generation of physically stable brick structures
  • Developers building or adapting tools for text-to-brick design
  • Brick-model designers who want prompt-generated concepts with LDraw-format output

Documented uses

  • Generate a toy-brick structure concept from a written description
  • Save a generated structure as a rendered PNG, brick-by-brick text, and LDraw model
  • Explore how stability analysis affects generated brick configurations
Before you choose

Requirements & limits

The repository is public and MIT-licensed. Local inference requires access to the gated Llama-3.2-1B-Instruct model and a Hugging Face token. A Gradio demo is linked from the repository.

  • Gurobi is optional but recommended for physics-based stability analysis; the documented alternative is simpler and less accurate.
  • Rendering through the inference script requires optional ImportLDraw and LDraw parts-library setup.
  • The CLI example saves a PNG preview, a brick-by-brick text file and an LDraw-format file.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Install BrickGPT for local inference

Set up the Python project and its documented model-access prerequisites for interactive text-to-brick inference.

Before you start

  • Python project manager uv
  • Access to the gated Llama-3.2-1B-Instruct model
  • A Hugging Face user access token set as HF_TOKEN
  • Gurobi license for physics-based stability analysis, or use the documented connectivity-based alternative

Steps

  1. Request access to Llama-3.2-1B-Instruct and create a Hugging Face user access token; set the token as the HF_TOKEN environment variable.

  2. Clone the BrickGPT repository and change into its directory.

  3. Install uv, then run the command below to create the project environment and install dependencies.

    Documented command
    uv sync

  4. Run the command below to start the interactive prompt session; use --use_gurobi False if running without Gurobi.

    Documented command
    uv run infer

Expected outcome

An interactive local inference session for entering text prompts and generating brick structures.

Guide limitations

  • The base Llama model is gated and requires access approval.
  • Without Gurobi, stability analysis uses a simpler, less accurate connectivity-based method.
  • The guide covers project installation and starting inference; rendering visualizations requires the additional ImportLDraw setup.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
16 October 2026
Original website title
AvaLovelace1/BrickGPT · Official homepage title · retained in its source language
How we review sourcesSuggest a correction