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

Zero Painter

Zero Painter is an open-source Python implementation of a training-free text-to-image method that combines a global scene prompt, object-specific descriptions and masks to control generated objects’ placement and shape.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Zero Painter — 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 Zero Painter when a text prompt alone does not provide enough control over where objects appear or what shapes they take, and you can supply mask and metadata inputs.

Who it is for

  • Developers integrating mask-conditioned image generation into a Python workflow.
  • Designers and creators who can prepare object masks and want more control over image composition.
  • Researchers exploring layout-conditional text-to-image synthesis.

Documented uses

  • Generate images from a global scene prompt and descriptions for individual objects.
  • Guide object placement and shape with masks.
  • Create mockups or scenes with separately described, positioned elements.
Before you choose

Requirements & limits

The repository is public and licensed under MIT. Its documented inference workflow uses Python, locally downloaded model files and command-line inputs.

  • The workflow depends on preparing a mask image and JSON metadata with a global prompt and object descriptions.
  • The method’s PACA and ReGCA attention stages are intended to align objects with descriptions and mask shapes.
  • The repository documents local script-based inference rather than a graphical interface or hosted generation service.
  • The repository lists Python as its primary language.

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

Run inference from masks and prompt metadata

Generate an image using the repository’s documented inference script and prepared mask-prompt inputs.

Before you start

  • Python environment with dependencies installed using pip3 install -r requirements.txt.
  • Downloaded Zero-Painter model files placed in the repository’s models folder.
  • A mask image and JSON metadata containing a global prompt and object descriptions.

Steps

  1. Install the requirements with the command below.

    Documented command
    pip3 install -r requirements.txt

  2. Download the models and place them in the models folder.

  3. Prepare a mask file and metadata JSON in the demonstrated format.

  4. Run the command below.

    Documented command
    python zero_painter.py --mask-path data/masks/1_rgb.png --metadata data/metadata/1.json --output-dir data/outputs/

Expected outcome

The script is provided for inference on the supplied mask and prompts pair, with an output directory specified for generated images.

Guide limitations

  • The repository does not document how to create masks or prepare the inputs beyond its example metadata format.
  • The instructions require downloaded model files; no hosted inference flow is described.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
18 October 2026
Original website title
Picsart-AI-Research/Zero-Painter · Official homepage title · retained in its source language
How we review sourcesSuggest a correction