Raw capture — raw/besser-pearl.md

url: https://besser-pearl.org/ title: BESSER author: LIST (Luxembourg Institute of Science and Technology) and SnT / University of Luxembourg date_fetched: 2026-07-25 date_published: unknown

BESSER

BESSER is an open-source low-code platform for building smart, AI-enhanced applications. It is led by LIST (Luxembourg Institute of Science and Technology) with participation from SnT / University of Luxembourg, funded by an FNR PEARL grant. The project involves 11 researchers and engineers, has produced 53 peer-reviewed papers, and participates in 2 EU Horizon projects. Licensed under MIT (v7.13.0).

The core paradigm is "model once, generate everywhere" — describe a system once in the platform's modeling environment, then generate working software across multiple technology stacks via 15+ code generators. Generators target Python, FastAPI, Django, SQLAlchemy, SQL, Java, React, and PyTorch.

Modeling Dimensions

BESSER supports multi-dimension modeling including: - Class diagrams (structural modeling) - State machines (behavioral modeling) - Object diagrams - Agent specifications - Deployment specifications

User Interfaces

• Free online visual editor at editor.besser-pearl.org
• Image-to-model (generate models from diagram images)
• Text-to-model via chatbot (natural language to models)
• No installation required for the online editor

Workflow

1. Model — describe the system in the web editor using diagrams, forms, or natural-language prompts
2. Generate — produce REST APIs, databases, back-ends, and AI agents via 15+ generators
3. Deploy — ship from the editor using generated deployment manifests

AI Integration

BESSER uses AI in two directions: - AI assists the modeling process (image-to-model, text-to-model) - Generated systems can embed AI components (PyTorch generator for ML models)

Tagline

"Better Software, Faster"

Funding

FNR PEARL grant (Luxembourg National Research Fund)