Python Ecosystem
FAI for Python
Two packages, one ecosystem. Offline-first SDK for knowledge access + MCP Server for AI agent integration.
Python SDK
pip install frootai · v3.3.0 · Zero dependencies
Install
pip install frootaiSolution plays offline
Knowledge modules
Cost estimation engine
A/B testing & evaluation
Comprehensive AI glossary terms
Zero external dependencies
Quick Start
from frootai import FrootAI
fai = FrootAI()
# Search knowledge
results = fai.search("RAG architecture")
# Get a solution play
play = fai.get_play("01-enterprise-rag")
# Estimate costs
cost = fai.estimate_cost("enterprise-rag", scale="prod")Python MCP Server
pip install frootai-mcp · v3.2.0 · Same MCP tools as Node
Install
pip install frootai-mcpMCP tools for any agent
Works with Claude, GPT, Gemini
Bundled knowledge (682KB)
Graceful offline fallback
Agent chain: build → review → tune
Architecture pattern advisor
MCP Config
{
"mcpServers": {
"frootai": {
"command": "frootai-mcp",
"args": []
}
}
}Which one do I need?
| Feature | Python SDK | Python MCP Server |
|---|---|---|
| Install | pip install frootai | pip install frootai-mcp |
| Use case | Your Python code calls it | AI agent calls it via MCP |
| Tools | Python API (functions) | MCP tools (protocol) |
| Dependencies | Zero | MCP SDK |
| Best for | Scripts, notebooks, apps | Claude, Copilot, Cursor agents |