Examples

Real-world examples showing both the deployment recipe (recipaie) and execution records (prodaie) together.

Deployment Examples

Each example shows the recipaie template and, where available, the prodaie execution record demonstrating real deployments.

nodejs-app

Deploy a Node.js application - supports PM2, systemd, Docker, or Kubernetes

Includes execution record

View Example

publog

Deploy Publog centralized log viewer - supports Docker Compose, Docker Swarm, Kubernetes, or standalone binary

Includes execution record

View Example

redis

Deploy Redis in-memory data store - supports Docker, Kubernetes, package manager, or source build

Includes execution record

View Example

tool-mcp

Install MCP bridge tools for AI assistant integration - standalone binaries, no containers required

Includes execution record

View Example

Creating Your Own Recipaie

To create a recipaie (deployment recipe) for your tool:

  1. Start with the specification
  2. List the prerequisites (docker, git, specific tools)
  3. Break deployment into atomic steps
  4. For each step, document: what, why, when, how, verify
  5. Add skip_if conditions for idempotent execution
  6. Be environment-agnostic - support Docker, Kubernetes, systemd, standalone binary, etc.
  7. Validate against the schema

How AI Creates a Prodaie

A prodaie is the execution record - the "product" of running a recipaie. It's typically created by an AI assistant as it works through the deployment.

AI Workflow

  1. Read the recipaie - AI receives the deployment recipe as context
  2. Initialize the prodaie - Create a location entry for this deployment
  3. Execute step by step - For each step:
    • Check skip_if condition - skip if already satisfied
    • Execute the "how" instruction
    • Run the "verify" check
    • Record: success: true/false and details: "what happened"
  4. Set location status - Mark as "complete", "partial", or "failed"
  5. Save the prodaie - Preserve for future reference

Step Result Fields

Each prodaie step includes two result fields:

FieldTypeDescription
successbooleantrue if step succeeded or was intentionally skipped
detailsstringHuman-readable explanation of what happened

Example Transformation

A recipaie step:

{
  "order": 3,
  "what": "Install or deploy Redis",
  "why": "Get Redis running in your environment",
  "when": "After prerequisites are ready",
  "skip_if": "Redis already running at target version",
  "how": "docker run -d --name redis redis:7-alpine",
  "verify": "redis-cli ping returns PONG"
}

Becomes a prodaie step after AI execution:

{
  "order": 3,
  "what": "Install or deploy Redis",
  "why": "Get Redis running in your environment",
  "when": "After prerequisites are ready",
  "skip_if": "Redis already running at target version",
  "how": "docker run -d --name redis redis:7-alpine",
  "verify": "redis-cli ping returns PONG",
  "success": true,
  "details": "Ran: docker run -d --name redis-dev -p 6379:6379 redis/redis-stack:latest. Verified: redis-cli ping returned PONG."
}