Skip to main content

MCP Server

The FrootAI MCP Server is a Model Context Protocol server that gives AI assistants access to the entire FrootAI knowledge base — 17 FROOT modules, 100 solution plays, 830+ primitives, and AI model comparison data.

Installation​

npx frootai-mcp@latest

No configuration needed — the server runs with bundled knowledge data.

Configuration​

VS Code (Copilot Chat)​

Add to .vscode/mcp.json in your project:

.vscode/mcp.json
{
  "servers": {
    "frootai": {
      "command": "npx",
      "args": ["frootai-mcp@latest"],
      "cwd": "${workspaceFolder}"
    }
  }
}

Cursor​

Add to your Cursor MCP settings:

~/.cursor/mcp.json
{
  "mcpServers": {
    "frootai": {
      "command": "npx",
      "args": ["frootai-mcp@latest"]
    }
  }
}

Claude Desktop​

Add to claude_desktop_config.json:

{
  "mcpServers": {
    "frootai": {
      "command": "npx",
      "args": ["frootai-mcp@latest"]
    }
  }
}

Available Tools​

Knowledge & Learning​

ToolDescription
search_knowledgeSearch across all 17 FROOT modules by topic
get_moduleGet full content of a specific module (F1–T3)
list_modulesList all modules organized by FROOT layer
lookup_termLook up an AI/ML term in the glossary (200+ terms)

Solution Plays​

ToolDescription
get_play_detailDetailed architecture, services, config for a play
list_community_playsList all 100 plays with status
semantic_search_playsDescribe what you want, get matching plays
compare_playsSide-by-side comparison of 2–3 plays
generate_architecture_diagramGenerate Mermaid.js diagram for a play

Models & Cost​

ToolDescription
compare_modelsSide-by-side model comparison for a use case
get_model_catalogList Azure OpenAI models with capabilities
estimate_costCalculate monthly Azure costs for a play
get_azure_pricingPricing estimates by scenario

Architecture & Patterns​

ToolDescription
get_architecture_patternGuidance for RAG, agents, cost optimization
get_froot_overviewComplete FROOT framework overview
get_github_agentic_osGitHub Copilot agentic OS guide

Build & Validate​

ToolDescription
agent_buildBuilder agent — implementation guidelines
agent_reviewReviewer agent — security + quality checklist
agent_tuneTuner agent — production readiness validation
validate_configValidate TuneKit config files
run_evaluationCheck AI quality scores against thresholds

Ecosystem​

ToolDescription
list_primitivesBrowse all 830+ primitives by type
fetch_azure_docsFetch latest Azure documentation
fetch_external_mcpSearch for external MCP servers
embedding_playgroundCompare texts for semantic similarity

Usage Examples​

Search for RAG patterns​

Ask your AI assistant:

"Search FrootAI knowledge for RAG chunking strategies"

The assistant calls search_knowledge with query: "RAG chunking strategies" and returns relevant sections from the FROOT knowledge base.

Find a solution play​

"I want to build a document processing pipeline. What FrootAI play should I use?"

The assistant calls semantic_search_plays to find the best match.

Compare models​

"Compare GPT-4o vs GPT-4o-mini for a RAG chatbot"

The assistant calls compare_models with useCase: "RAG chatbot".

Estimate costs​

"How much would Play 01 cost in production?"

The assistant calls estimate_cost with play: "01" and scale: "prod".

Docker Deployment​

For containerized environments, see the Docker distribution.

Version​

The MCP server version is synced with the main FrootAI release. Current: v3.5.0.

# Check installed version
npx frootai-mcp@latest --version

See Also​