AIDC-AI.IO — MCP Connector

AI data center sizing, validation, and layout via a remote MCP server.
AI 데이터센터 자동화 툴: 결정론적 엔진으로 AI 데이터센터를 설계·검증·레이아웃합니다.

---

What is this?

This repository shows how to connect an MCP client or REST client to the AIDC-AI.IO Design Engine — a deterministic, source-backed engine that sizes, validates, and lays out Rubin-era AI data centers.

What the engine does (on the server):

• Accepts an IT load, rack density, GPU generation (Hopper / Blackwell / NVIDIA Vera Rubin NVL72 / VR200), and site constraints.
• Returns deployment-unit-snapped rack counts, design PUE, power-factor-backed total MVA (22.9 kV intake), liquid-cooling / air-cooling heat split, CDU planning values, cost (KRW), and timeline.
• Validates designs against electrical, cooling, layout, safety, and data rules with severity-classified findings and RFIs.
• Generates a rack-plan grid (hall dimensions, row/column positions in mm) and a site-block layout.

What this repo contains:

• MCP client configuration snippet.
• curl and Node.js examples that call the public REST projection (/api/agent/*).
• An illustrative response so you know what fields to expect.

The core calculation engine, reference catalogs (rack library, AHJ/code matrix, 1.6T fabric topology, direct-to-chip (D2C) cooling models, etc.) are proprietary and remain server-side. No engine source is published here.

Korea live. Region-specific: 22.9 kV utility intake, Korean AHJ/code, climate, and operations validation. Keywords the engine targets: AI data center, AIDC, NVIDIA Rubin, Vera Rubin, 22.9kV, liquid cooling, CDU, D2C, 1.6T fabric, PUE.

---

MCP Server

Field	Value
Transport	Streamable HTTP
Endpoint	https://aidc-ai.io/api/mcp
Official registry name	io.aidc-ai/design-engine
Auth	None required (anonymous tier). Optional Authorization: Bearer aidc_live_<32hex> raises rate tier.
Tool count	3
Rate limit (anon)	10 req / hour on /api/agent/*

Tools

Tool	One-line description
design	Size an AI data center: returns rack count, PUE, total MVA, liquid/air cooling split, CDU count, cost (KRW), and build timeline.
validate	Check a design against electrical, cooling, layout, safety, and data rules; returns severity-classified findings and RFIs.
layout	Generate a rack-plan grid (hall dimensions, row/column positions in mm) and a site-block layout.

---

Quick Start

MCP client configuration

Add this to your MCP client config (e.g. Claude Desktop claude_desktop_config.json, Cursor MCP settings, or any Streamable HTTP client):

{
  "mcpServers": {
    "aidc-design-engine": {
      "url": "https://aidc-ai.io/api/mcp"
    }
  }
}

The server is immediately usable without an API key. To raise the rate limit, add:

{
  "mcpServers": {
    "aidc-design-engine": {
      "url": "https://aidc-ai.io/api/mcp",
      "headers": {
        "Authorization": "Bearer aidc_live_<your-32-hex-key>"
      }
    }
  }
}

Contact contact@aidc-ai.io for a registered or partner key.

Docker (local stdio server)

Build and run the same published MCP server used for registry evaluation:

docker build -t aidc-ai-mcp .
docker run --rm -i aidc-ai-mcp

The container communicates over stdio and connects to https://aidc-ai.io by default. No API key is required for the anonymous tier.

---

REST Usage

The MCP tools proxy to these REST endpoints (permissive CORS, same optional auth):

Tool	REST endpoint
design	POST https://aidc-ai.io/api/agent/design
validate	POST https://aidc-ai.io/api/agent/validate
layout	POST https://aidc-ai.io/api/agent/layout

Example: size a 30 MW Rubin-era AI data center

curl -s -X POST https://aidc-ai.io/api/agent/design \
  -H "Content-Type: application/json" \
  -d '{
    "itLoadMw": 30,
    "rackDensityKw": 120,
    "gpuGen": "rubin",
    "siteAreaSqm": 5000,
    "region": "metropolitan",
    "options": {
      "redundancy": "n_plus_1",
      "coolingMode": "liquid",
      "pueTarget": 1.2
    }
  }'

Illustrative response

The JSON below is illustrative — field names and structure reflect the actual API shape, but exact numbers will vary by engine version and input. See design.response.example.json for the full object.

{
  "rackCount": 256,
  "rackCountRaw": 250,
  "pueDesign": 1.21,
  "mvaTotal": 45.8,
  "liquidCoolingLoadMw": 26.4,
  "airCoolingLoadMw": 3.6,
  "cduCount": 13,
  "totalCostKrw": 187500000000,
  "totalMonths": 28,
  "warnings": []
}

(30 MW IT / 120 kW per rack / Rubin / 5 000 m² / metropolitan / N+1 / liquid / PUE 1.2 target)

---

Tools — Input Reference

design

Size an AI data center from scratch.

Field	Type	Range / values	Required
itLoadMw	number	0 < x ≤ 1000	Yes
rackDensityKw	number	0 < x ≤ 500	Yes
gpuGen	string	"hopper" | "blackwell" | "rubin"	Yes
siteAreaSqm	number	0 < x ≤ 1 000 000	Yes
region	string	"metropolitan" | "regional"	Yes
options.redundancy	string	"n" | "n_plus_1" | "2n"	No
options.coolingMode	string	"air" | "hybrid" | "liquid"	No
options.pueTarget	number	1.0 – 2.5	No

Key response fields: rackCount, rackCountRaw, pueDesign, mvaTotal, liquidCoolingLoadMw, airCoolingLoadMw, cduCount, totalCostKrw, totalMonths, warnings[]

---

validate

Check a design against engineering rules.

{
  "rawInput": {
    "itLoadMw": 30,
    "rackDensityKw": 120,
    "gpuGen": "rubin",
    "siteAreaSqm": 5000,
    "region": "metropolitan"
  }
}

Key response fields: findings[] (each with severity, code, message), rfis[], passCount, warnCount, failCount

---

layout

Generate a rack plan and site block layout.

{
  "design": {
    "itLoadMw": 30,
    "rackDensityKw": 120,
    "gpuGen": "rubin",
    "siteAreaSqm": 5000,
    "region": "metropolitan"
  },
  "siteCentroid": { "lat": 37.5665, "lng": 126.9780 },
  "siteAreaSqm": 5000
}

Key response fields: rackPlan (hall dimensions, rows, columns, per-rack positions in mm), sitePlan (block-level layout in percentage coords)

---

Links

Resource	URL
Website	https://aidc-ai.io
OpenAPI 3.1 spec	https://aidc-ai.io/api/openapi.json
MCP server card	https://aidc-ai.io/.well-known/mcp/server.json
LLM context	https://aidc-ai.io/llms.txt
Full LLM context	https://aidc-ai.io/llms-full.txt
Contact	contact@aidc-ai.io

---

License

This repository (examples and connector code only) is released under the MIT License.
The AIDC-AI.IO engine, reference catalogs, and all server-side logic remain proprietary.