MCP Heuristics - Garden Management System
About
POC for tracking MCP server tool usage with Kestra
Details
- Author
- allenheltondev
- GitHub stars
- 1
- Downloads
- 246
- Categories
- Other
Jump to
- Garden bed management with metadata and updates
- Plant management with planting dates and species
- Observation logging for moisture, pests, and health
- Harvest tracking with quantities and dates
- Automatic telemetry for tool usage and performance
Install Node.js, PostgreSQL, and the Kestra CLI. Create a .env file with DATABASE_URL (Neon PostgreSQL) and MOMENTO_TOKEN, then install dependencies.
MCP Heuristics - Garden Management System
A Model Context Protocol (MCP) server for tracking and managing permaculture garden beds, built with TypeScript and PostgreSQL.
Overview
This system provides a comprehensive suite of tools for managing garden beds, plants, observations, and harvests. It uses the Model Context Protocol to provide structured interactions with the garden management system.
Features
- Garden Bed Management
- Add new garden beds with metadata
- Update existing garden bed information
- List all garden beds
- Get detailed information about specific beds
- Plant Management
- Add plants to garden beds
- Remove plants from beds
- Track planting dates and species
- Observation System
- Log observations about garden beds
- Track moisture levels, pest presence, and plant health
- View historical observations
- Harvest Tracking
- Log harvest information
- Track harvest quantities and dates
- View harvest history
- Telemetry
- Automatic tracking of tool usage
- Performance metrics collection
- Success/failure monitoring
Technical Stack
- TypeScript
- PostgreSQL (via Neon Database)
- Model Context Protocol SDK
- Kestra (for workflow automation)
- Momento Cache
Prerequisites
- Node.js
- PostgreSQL
- Kestra CLI (for running workflows)
Environment Setup
1. Create a .env file with the following variables:
DATABASE_URL=your_neon_database_url
MOMENTO_TOKEN=your_momento_token
2. Install dependencies
Sign in to leave a review
Use Google, GitHub, or an email account so ratings stay tied to real people.
No reviews posted yet.



