Welcome to Pili’s Documentation!

Pili is a sophisticated multiagent chatbot microservice for exercise tracking, built with FastAPI, LangGraph, and LangSmith. It features a 3-agent orchestration architecture with intelligent routing and MCP server integration for the Scaffold Your Shape fitness platform.

🏗️ Architecture Overview

Pili implements a cutting-edge 3-agent orchestration system:

• 🎯 Orchestration Agent: Intelligent request routing and response coordination

• 📝 Logger Agent: Direct MCP server integration for activity tracking

• 🏃‍♀️ Coach Agent: AI-powered workout planning and progress analysis

🚀 Key Features

• Intelligent Request Routing: Advanced NLP-based intent detection

• Real-time Streaming: FastAPI streaming responses with orchestration metadata

• Conversation Memory: Persistent context across user interactions

• MCP Server Integration: Direct connection to Scaffold Your Shape platform

• Robust Error Handling: Graceful fallbacks for all components

• Auto-generated API Documentation: Interactive Swagger UI

📚 Documentation Structure

Getting Started

• Installation Guide
  • Prerequisites
  • System Requirements
  • Quick Installation
  • Environment Configuration
  • Docker Installation
  • Development Setup
  • Troubleshooting
  • Next Steps
• Quick Start Guide
• 🚀 5-Minute Setup
• 🎯 First Interactions
• 📱 Using the Interactive API
• 🧠 Understanding Pili’s Responses
• 🔧 Configuration Check
• 🎭 Agent System Demo
• 🔌 MCP Server Integration
• 🚨 Common Issues & Solutions
• 🎉 Next Steps
• 💡 Pro Tips
• Configuration Guide
• Environment Variables
  • Core Configuration
  • LLM Provider Configuration
  • MCP Server Configuration
  • Memory Configuration
  • Application Configuration
• Configuration Files
  • Settings Module
  • Agent Configuration
• Development Configuration
  • Debug Mode
  • Testing Configuration
• Production Configuration
  • Security Settings
  • Performance Settings
  • Monitoring
• Docker Configuration
  • Docker Compose
  • Environment File
• Configuration Validation
• Configuration Templates
  • Development Template
  • Production Template
  • Testing Template

Architecture

• Architecture Overview
• System Architecture
• Core Components
• Agent Interaction Patterns
  • Simple Request Flow
  • Complex Request Flow
• Request Routing Logic
• Technology Stack
• Data Flow Architecture
• Scalability Considerations
• Security Architecture
• Deployment Architecture
• Future Architecture Considerations

Examples

• Basic Usage Examples
• Getting Started
• Basic Examples
• API Endpoints
• Next Steps

🤖 What Can Pili Do?

Category	Example Messages	Agent Flow
Activity Logging	“I ran 5 km”, “Did yoga for 45 minutes”	Orchestration → Logger → MCP Server
Club Management	“Show me clubs”, “Join club runners”	Orchestration → Logger → MCP Server
Workout Planning	“Create a running plan”, “I need a training schedule”	Orchestration → Logger (data) + Coach (planning)
Progress Analysis	“How am I doing?”, “Analyze my progress”	Orchestration → Logger (data) + Coach (analysis)
Motivation	“I need motivation”, “Encourage me”	Orchestration → Coach with context

🔗 Quick Links

• GitHub Repository: Scaffold-your-shape-Chatbot

• API Documentation: http://localhost:8000/api/docs

• LangSmith Dashboard: Monitor Conversations

• Scaffold Your Shape: Main Platform

📧 Support

If you encounter any issues or have questions:

• Check the installation guide

• Review the quickstart section

• Open an issue on GitHub

• Contact the development team

Indices and tables

• Index

• Module Index

• Search Page

—

Meet Pili - Your friendly AI fitness companion that makes tracking workouts as easy as having a conversation! 🏃‍♀️💪