Shipped: Meal Magic — A Local-First Cooking Skill for AI Agents

Meal Magic is an open-source, local-first cooking skill that gives AI agents persistent memory for everyday meal planning.

Instead of starting from scratch in every conversation, it helps AI assistants remember household preferences, dietary restrictions, favorite recipes, and what people actually cooked or ate.

Key features include:

• Personalized, step-by-step recipes based on available ingredients.
• Household and member-specific food preferences.
• Persistent meal history, with separate cooked and eaten events.
• Recipe management, favorites, and multi-day meal plans with shopping lists.
• Meal variety and frequency statistics.
• Shared local SQLite storage across supported AI agent environments.

Built with Python and SQLite, Meal Magic works with Codex, Antigravity, and Claude Code through a lightweight agent skill. It requires no additional cloud service, API key, or subscription beyond the user's existing AI agent setup.

Version 0.7.0 introduces complete recipe management, improved search, meal history filtering, and performance optimizations.

The goal is simple: make AI cooking assistants genuinely useful by helping them remember, adapt, and improve everyday meal decisions.