MCP Server Reference

The AIDH MCP server gives AI agents access to the full AI Data Hub directory. Install it locally via STDIO or connect remotely via SSE.

Installation

Option 1: npx (recommended)

>npx @aidh/mcp-server

Option 2: Claude Desktop config

{
  "mcpServers": {
    "aidh": {
      "command": "npx",
      "args": ["-y", "@aidh/mcp-server"]
    }
  }
}

Option 3: Global install

>npm install -g @aidh/mcp-server && aidh-mcp

Available Tools (10)

search_tools

Search AI tools by category, vertical, budget, and features.

Parameters:
  query?: string       # Free-text search
  category?: string    # e.g., "coding", "marketing"
  vertical?: string    # Tool vertical filter
  maxPrice?: number    # Max starting price (USD)
  hasApi?: boolean     # Filter for API access
  limit?: number       # Max results (default: 10)

get_tool_details

Get full details on a specific AI tool by slug.

Parameters:
  slug: string         # e.g., "cursor", "chatgpt"

compare_tools

Side-by-side comparison of 2+ tools.

Parameters:
  slugs: string[]      # 2-5 tool slugs to compare

compare_llm_pricing

Compare LLM model pricing and capabilities.

Parameters:
  provider?: string    # e.g., "Anthropic", "OpenAI"
  category?: string    # frontier, mid-tier, budget, etc.
  multimodal?: boolean # Filter multimodal models
  openSource?: boolean # Filter open source models

recommend_model

Get ranked model recommendations for a task.

Parameters:
  task: string         # Describe your task
  budget?: string      # low, medium, high, unlimited
  requireMultimodal?: boolean
  requireOpenSource?: boolean
  minContextWindow?: number

estimate_cost

Calculate API costs for a workload.

Parameters:
  model?: string       # Model slug (omit for comparison)
  dailyInputTokens: number
  dailyOutputTokens: number
  daysPerMonth?: number

search_datasets

Find AI/ML datasets by type, license, and format.

Parameters:
  query?: string
  category?: string    # text, code, image, audio, etc.
  commercialUse?: boolean
  format?: string      # JSON, Parquet, CSV, etc.

search_mcp_servers

Find MCP servers by category and transport.

Parameters:
  query?: string
  category?: string    # development, data, etc.
  transport?: string   # stdio, sse
  setupComplexity?: string

check_agent_readiness

Audit a URL for agent-friendliness.

Parameters:
  url: string          # URL to audit

get_glossary_term

Look up AI/ML terminology definitions.

Parameters:
  term: string         # e.g., "RAG", "MCP", "fine-tuning"