Zonatic

Zonatic is an experimental geospatial rule engine.

It focuses on a simple but often overlooked problem: how geographic boundaries are used to make explicit, deterministic decisions — not just drawn on a map.

This repository provides a hosted sandbox playground where anyone can:

• a sandbox API for evaluating boundary-based decisions
• an interactive sandbox UI for authoring and inspecting boundaries

⚠️ Zonatic is a playground and exploration project, not a production-ready service.

---

What Zonatic is (and is not)

Zonatic is:

• a boundary-based decision engine
• deterministic and auditable
• API-first and agent-friendly
• focused on rules, not UI

Zonatic is NOT:

• a tracking application
• a full SaaS platform
• a realtime GPS system
• an AI decision maker

AI (or LLMs) can control Zonatic, but never decide — the core logic remains explicit and deterministic.

---

Sandbox Concept

Zonatic revolves around the idea of a sandbox.

A sandbox represents an isolated, temporary environment where:

• a geographic boundary is defined (GeoJSON)
• location checks are evaluated against that boundary
• decisions are deterministic and explainable
• state expires automatically

The sandbox exists in two forms:

• Sandbox API — for programmatic evaluation
• Sandbox UI — for interactive authoring and inspection

The sandbox UI lives in the sandbox/ directory.

---

Try Zonatic in 2 Minutes (API)

1. Open Swagger UI

http://localhost:8080/docs/

This is the fastest way to try Zonatic without writing any code.

---

2. Create a Sandbox

Call:

POST /sandbox

Response example:

{
  "sandbox_id": "sbx_123abc",
  "token": "sk_zonatic_xxxxx",
  "endpoint": "/sandbox/sbx_123abc/check",
  "expires_at": "2025-01-01T12:00:00Z"
}

Each sandbox:

• is isolated
• supports one boundary
• expires automatically

---

3. Save Your Boundary

Call:

PUT /sandbox/{id}/boundary

Body example (GeoJSON Polygon):

{
  "type": "Feature",
  "geometry": {
    "type": "Polygon",
    "coordinates": [
      [
        [106.8, -6.2],
        [106.81, -6.2],
        [106.81, -6.21],
        [106.8, -6.21],
        [106.8, -6.2]
      ]
    ]
  }
}

---

4. Check a Location

Call:

POST /sandbox/{id}/check

Body:

{
  "lat": -6.2,
  "lng": 106.8
}

Response:

{
  "status": "INSIDE",
  "explanation": "Location is inside the defined boundary"
}

This result is deterministic and explicit and explainable.

---

Using Zonatic with ChatGPT (Conceptual)

Zonatic is designed to be agent-friendly.

A common setup:

• Zonatic exposes a REST API
• ChatGPT (via Custom GPT Actions) calls the API
• ChatGPT explains results, but does not make decisions

Example prompt:

Check whether this coordinate is inside my boundary.

Zonatic:

• evaluates the rule
• returns a structured decision

ChatGPT:

• translates the result into natural language

This separation keeps the system auditable and safe.

---

API Reference

The full API contract is defined in:

GET /openapi.yaml

Interactive documentation:

/docs/

Project Status

• Core sandbox API: ✅
• Boundary evaluation: ✅
• Sandbox UI: ✅ (experimental)
• Swagger UI: ✅
• Hosting: experimental
• Persistence: in-memory (v0)

Planned explorations:

• group / collective presence checks
• expected absence / readiness rules
• MCP-compatible interfaces
• optional AI-assisted authoring (non-decision)

---

LICENSE

Apache License 2.0

---

Why this exists

Zonatic started as an exploration after building a small web component for drawing boundaries. The question that followed was simple:

After a boundary is defined — what decisions does it actually enable?

This repository is an attempt to answer that question, openly and incrementally.