StrainsDB: AI-Generated Cannabis Database

An experimental Django application exploring AI-powered content generation

strainsdb.org is a Django web application that catalogues various cannabis strains, built as a proof of concept for AI-generated web contentThis project represents an early exploration into AI-powered content generation for web applications, predating the widespread adoption of large language models for automated content creation..

Features

The application provides:

• Strain Browser: Browse through different cannabis strains with detailed information
• Search Functionality: Find specific strains based on characteristics and effects
• Terpene Profiles: Detailed terpene information for each strain
• AI-Generated Content: All strain descriptions and data generated using GPT-3.5

Technical Architecture

The backend leverages modern Python tooling:

• Django Framework: Provides the web application structure and admin interface
• GPT-3.5 Integration: Generates strain descriptions and terpene profiles
• Instructor Library: Uses instructorInstructor is a Python library that uses Pydantic models to structure and validate LLM outputs, ensuring reliable data extraction from language models for programmatic use. to structure and validate AI outputs with Pydantic models
• SQLite Database: Stores all strain and terpene dataSQLite provides a lightweight, serverless database solution perfect for proof-of-concept applications, requiring no additional infrastructure while maintaining ACID compliance.

Innovation

StrainsDB demonstrates how AI can be integrated into traditional web applications to generate rich, structured content. The use of Pydantic models ensures the AI-generated data maintains consistency and validity, making it suitable for production use.

Visit: strainsdb.org