About this project

ERP-CRM is a project that combines classic business management with an artificial intelligence component for 3D product rendering. According to its documentation, the platform is organized into CRM modules (customers, contacts, and sales opportunities), inventory (stock, products, and warehouses), billing (invoices, proformas, and tax management), projects with associated tasks, 3D render, and a "digitizer" module oriented toward the catalog. Architecture: it follows a client-server scheme with separation of responsibilities. The frontend is a JavaScript application (HTML and CSS) providing access to the different modules; the backend is a Python API, with specific files for the AI engine (ai_engine.py) and the 3D rendering logic (render_3d.py). Deployment is supported by Docker and Docker Compose, and the repository includes GitHub Actions workflows along with Dependabot configuration. Installation: the README proposes two paths. Local installation requires Node.js 18+ with npm, Python 3.10+ and pip; frontend and backend dependencies are installed, .env.example is copied to .env, and the frontend (npm start) and backend (python app.py) are started separately. The container alternative is reduced to docker-compose up --build. Environment variables: the template includes database parameters (DB_HOST, DB_PORT, DB_NAME, DB_USER, DB_PASSWORD), AI provider credentials for rendering (AI_API_KEY, AI_PROVIDER), and server settings (PORT, FLASK_ENV, SECRET_KEY). The README itself warns that the .env file must not be uploaded to the repository. Usage: the frontend is accessed at http://localhost:3000 and the API is available at http://localhost:5000, with configurable ports. 3D renders are generated from the Digitizer module by configuring product parameters. Testing: npm test for the frontend and pytest for the backend are documented, with automatic execution on every push and pull request via GitHub Actions. The indicated license is MIT. Documentation scope: the README describes the architecture, modules, and startup process, but does not detail the data model, specific API endpoints, or the internal workings of the AI engine, of which it only indicates that it depends on an external provider via API key. To assess the actual state of each module, it is advisable to review the source code before considering production use.