Guardian Engine — API & MCP Integration Guide

Deterministic verification infrastructure for AI agent outputs. Guardian Engine catches hallucinated temperatures, missing techniques, wrong ingredients, and impossible cooking steps before they reach the pan. Recipes are the first vertical — the same deterministic approach generalises to any procedural domain where correctness matters.

Endpoint: https://api.kaimeilabs.dev/mcp
Transport: Streamable HTTP (MCP)
Auth: None — free during early access (fair use applies)

---

Connect Your Agent

Guardian is a hosted MCP server. No install, no API key, no Docker. Pick your client and paste the config.

Claude Desktop

Add to your claude_desktop_config.json:

{
  "mcpServers": {
    "guardian": {
      "url": "https://api.kaimeilabs.dev/mcp",
      "transport": "streamable-http"
    }
  }
}

Restart Claude Desktop. Ask: "List the available dishes in Guardian Engine" to confirm.

Cursor

Open Settings → MCP Servers → Add new MCP server, then paste:

{
  "guardian": {
    "url": "https://api.kaimeilabs.dev/mcp",
    "transport": "streamable-http"
  }
}

VS Code (GitHub Copilot)

Add to your .vscode/mcp.json (or user settings.json under "mcp"):

{
  "servers": {
    "guardian": {
      "type": "http",
      "url": "https://api.kaimeilabs.dev/mcp"
    }
  }
}

Windsurf

Add to your Windsurf MCP config:

{
  "mcpServers": {
    "guardian": {
      "serverUrl": "https://api.kaimeilabs.dev/mcp"
    }
  }
}

Smithery (One-Click)

— auto-configures Claude Desktop, Cursor, and more.

[!WARNING] Smithery Proxy Limitation: The default Smithery proxy URL (guardian-engine--kaimeilabs.run.tools) does not support Streaming HTTP and will silently fail. You MUST edit your MCP config after installation to use the direct endpoint: https://api.kaimeilabs.dev/mcp.

Glama.ai

Guardian Engine is also listed on Glama.ai — discover and connect to MCP servers from the Glama directory.

Any MCP Client (Python SDK)

import asyncio
from mcp.client.session import ClientSession
from mcp.client.streamable_http import streamable_http_client
from httpx import AsyncClient

async def main():
    async with AsyncClient(timeout=30.0) as http:
        async with streamable_http_client("https://api.kaimeilabs.dev/mcp", http_client=http) as streams:
            read_stream, write_stream, _ = streams
            async with ClientSession(read_stream, write_stream) as session:
                await session.initialize()
                result = await session.call_tool("list_dishes", arguments={"cuisine_filter": "french"})
                print(result)

asyncio.run(main())

pip install mcp>=1.2.1 httpx>=0.27.0

---

Tools

verify_recipe

Verify a candidate recipe against a Guardian master recipe. Returns a structured report with verdict, score, and detailed findings.

Parameter	Type	Required	Description
dish	string	Yes	Name or alias of the dish (e.g. "carbonara", "rendang", "kung-pao", "bourguignon")
candidate_json	string	Yes	Full recipe as a JSON string — see schema.md
original_prompt	string	No	The user's original request that generated the recipe

Tip — pass the prompt for better feedback: When you include original_prompt (e.g. "Make a spicy vegan rendang"), Guardian activates Guided Oracle Mode: it reads the user's intent and returns specific, actionable improvement hints tailored to their request. Without it, Guardian returns only a Pass/Fail verdict and score.

list_dishes

List all master recipes Guardian can verify against.

Parameter	Type	Required	Description
cuisine_filter	string	No	Filter by cuisine (e.g. "french", "chinese", "thai")

---

Available Recipes (51 dishes, 4 continents)

Region	Dishes
Europe	Beef Bourguignon · Beef Wellington · Basque Cheesecake · Cacio e Pepe · Cheese Soufflé · Chocolate Soufflé · Confit de Canard · Coq au Vin · Crème Brûlée · Fish & Chips · Florentine Biscuits · French Onion Soup · Gazpacho · Gnocchi di Patate · Goulash · Osso Buco · Pasta alla Norma · Pasta Carbonara · Patatas Bravas · Ratatouille · Risotto alla Milanese · Roast Chicken · Spaghetti Aglio e Olio · Spaghetti Bolognese · Spanish Paella · Tiramisu · Tortilla de Patatas
Asia & Middle East	Beef Rendang · Biryani · Butter Chicken · Cantonese Steamed Fish · Chicken Tikka Masala · Falafel · Hummus · Kimchi Fried Rice · Kung Pao Chicken · Laksa · Nasi Goreng · Pad Thai · Tabbouleh · Thai Green Curry · Tonkotsu Ramen
Americas	Caesar Salad · Ceviche · Chili con Carne · Macaroni and Cheese · Mole Poblano · Southern Fried Chicken · Texas Smoked Brisket
Africa	Moroccan Lamb Tagine · Shakshuka

All recipes accept multiple aliases (e.g. "gong-bao", "tikka-masala", "risotto", "bourguignon"). Use list_dishes for the full live catalog.

Missing a Dish?

The catalog is regularly expanding. If your agent requires verification for a dish not currently supported, please open an issue on GitHub to request it. We prioritize additions based on developer demand.

---

Example Verification Output

What does a Guardian verification report actually look like? Here's the response structure when an AI agent submits a recipe with authenticity issues:

{
  "verdict": "FAILED",
  "authenticity_score": 72.4,
  "findings": [
    {
      "issue": "MISSING_REQUIRED_INGREDIENT",
      "severity": "CRITICAL",
      "justification": "This ingredient provides a signature flavour component essential to the dish's identity."
    },
    {
      "issue": "WRONG_COOKING_MEDIUM",
      "severity": "WARNING",
      "justification": "Cooking medium fundamentally affects texture and flavour."
    }
  ],
  "allergen_warnings": ["milk", "eggs"],
  "summary": {"INFO": 1, "WARNING": 1, "CRITICAL": 2}
}

Each finding includes a severity and a justification grounded in culinary science — letting the agent fix only what's wrong instead of guessing.

---

Files in This Repository

File	Purpose
schema.md	Complete candidate_json structure required by verify_recipe
client.py	Python example: submit a recipe for verification
test_integration.py	Live connectivity test against the public API
smithery.yaml	Smithery MCP registry configuration
glama.json	Glama.ai MCP server claim configuration

---

Data & Privacy

• No PII collected — we do not store user names, emails, or API keys. Underlying cloud infrastructure may temporarily process IP addresses for routing.
• Data for Compute Exchange — the free service is provided in exchange for usage data. Submitted recipes are used to improve verification accuracy and create anonymized derived datasets. See our Terms of Service.
• Do not include PII in recipe payloads.
• Fair use quotas enforced via compute limits.

[!CAUTION] Not a Substitute for Food Safety Knowledge
While Guardian Engine catches explicitly dangerous AI hallucinations (like serving poultry below safe temperatures), it cannot guarantee a recipe is 100% safe to consume. Pathogen destruction relies on variables (time, mass, equipment) that text-based AI models cannot perfectly control. Verification results are informational and must always be paired with human common sense and standard kitchen safety practices.

---

Support & Contact

Building an AI cooking assistant, smart kitchen platform, or agentic food-tech product? We'd love to hear from you.

• Email: partners@kaimeilabs.dev
• Website: kaimeilabs.dev
• GitHub: github.com/kaimeilabs

License

Client code in this repository (client.py, test_integration.py) is released under the MIT License. The Guardian Engine verification logic and master recipe datasets are proprietary.