🚀 PaceBuddy — Multi-Agent AI System for Task, Schedule & Information Management

An API-first, production-ready multi-agent AI system that intelligently manages tasks, schedules, reminders, and contextual information by orchestrating multiple specialized agents and external tools.

🌐 Live Demo: https://pacebuddy2-417661527307.us-central1.run.app/dev-ui/?app=PaceBuddy 📦 Repository: https://github.com/ashu485761/Pace-Buddy.git

---

📌 Problem Statement

Build a multi-agent AI system that helps users manage tasks, schedules, and information by interacting with multiple tools and data sources.

✅ Core Requirements Covered

• ✔ Primary orchestrator agent coordinating sub-agents
• ✔ Multi-agent workflow execution
• ✔ Task & schedule management
• ✔ Structured data handling
• ✔ API-based cloud deployment
• ✔ Tool integration support (calendar, reminders, notes, info retrieval)
• ✔ Scalable cloud-native architecture

---

🎯 What Makes PaceBuddy Special

PaceBuddy is not just a chatbot — it is a real-world productivity intelligence platform.

It uses a hierarchical multi-agent architecture where a main coordinator agent delegates responsibilities to domain-specific sub-agents.

This makes the system:

• ⚡ Faster in decision making
• 🧠 Better at task decomposition
• 🔄 Reliable in multi-step workflows
• ☁ Cloud deployable as microservice APIs
• 📈 Easy to scale with more tools and agents

---

🏗️ System Architecture

                        ┌─────────────────────┐
                        │     User Query      │
                        └──────────┬──────────┘
                                   │
                                   ▼
                        ┌─────────────────────┐
                        │  Primary AI Agent   │
                        │   (Orchestrator)    │
                        └───────┬─────┬───────┘
                                │     │
                ┌───────────────┘     └───────────────┐
                ▼                                     ▼
      ┌──────────────────┐                  ┌──────────────────┐
      │ Scheduling Agent │                  │ Information Agent│
      │ - reminders      │                  │ - search notes   │
      │ - events         │                  │ - retrieve data  │
      └──────────────────┘                  └──────────────────┘
                │                                     │
                └───────────────┬─────────────────────┘
                                ▼
                      ┌────────────────────┐
                      │ External MCP Tools │
                      │ Calendar / Notes   │
                      │ Task Storage / API │
                      └────────────────────┘

---

⚙️ Multi-Agent Workflow

The intelligence of PaceBuddy lies in how agents collaborate.

1) 🧠 Primary Orchestrator Agent

The main agent receives the user intent and performs:

• Intent detection
• Task classification
• Context understanding
• Agent delegation
• Final response synthesis

2) 📅 Scheduling Agent

Handles:

• Meeting reminders
• Deadlines
• Study schedules
• Habit tracking
• Daily planning

3) 📚 Information Agent

Handles:

• Retrieving saved notes
• Searching structured data
• Summarizing stored information
• Querying tool responses

4) 🔗 Tool Interaction Layer

The agents connect with:

• Calendar systems
• Reminder workflows
• Database / storage
• External APIs
• MCP-compatible tools

---

🔥 Key Features

✅ Intelligent Task Planning

Automatically converts natural language into executable tasks.

Example:

“Remind me to revise DSA tomorrow at 8 PM and save notes from today’s lecture.”

The orchestrator splits this into:

• Reminder creation
• Note storage
• Future retrieval reference

---

✅ Multi-Step Execution

Supports chained workflows such as:

1. Create task
2. Schedule reminder
3. Save metadata
4. Retrieve when requested
5. Update status

---

✅ Cloud-Native API Deployment

Deployed on Google Cloud Run, enabling:

• High availability
• Low-latency API responses
• Containerized deployment
• Production-ready endpoints

---

✅ Developer UI for Testing

The included Dev UI endpoint allows judges to directly test:

• agent orchestration
• workflow execution
• task routing
• session memory
• multi-turn conversations

---

🛠️ Tech Stack

💻 Backend

• Python
• FastAPI / Agent API
• Multi-Agent orchestration logic

☁ Cloud

• Google Cloud Run
• Service Account authentication
• Container Registry
• Environment-based config

🧠 AI Layer

• LLM-powered orchestration
• Tool routing
• Context memory
• Workflow decomposition

🔗 Integrations

• MCP tool connectors
• Reminder systems
• Calendar APIs
• Structured storage

---

📂 Project Structure

PaceBuddy/
├── agent.py
├── __init__.py
├── requirements.txt
├── .env
└── deployment configs

---

🚀 Deployment

Live Endpoint

https://pacebuddy2-417661527307.us-central1.run.app

Dev UI

https://pacebuddy2-417661527307.us-central1.run.app/dev-ui/?app=PaceBuddy

---

🧪 Example Test Cases for Judges

📌 Test 1 — Reminder Workflow

Prompt:

Remind me to submit my hackathon PPT tomorrow at 10 AM

Expected:

• task parsed
• schedule extracted
• reminder workflow triggered

📌 Test 2 — Multi-Agent Delegation

Prompt:

Save today’s thermodynamics notes and remind me to revise on Sunday

Expected:

• info agent stores note
• scheduler agent sets reminder
• orchestrator merges result

📌 Test 3 — Information Retrieval

Prompt:

What notes did I save for my DSA interview preparation?

Expected:

• retrieval agent searches stored data
• summarized response returned

---

🏆 Innovation Highlights

Why this stands out for judges:

• 🌟 Real-world usability beyond a simple chatbot
• 🌟 True multi-agent orchestration design
• 🌟 Production deployment on cloud
• 🌟 Scalable architecture for future tools
• 🌟 MCP-compatible workflow expansion
• 🌟 Practical student productivity use case

---

📈 Future Scope

• Google Calendar sync
• WhatsApp reminders
• Voice assistant support
• Team collaboration tasks
• Hackathon project planner
• Study + fitness schedule fusion
• Smart prioritization engine

---

👨‍💻 Author

Ashutosh Biswal B.Tech Student | AI + Multi-Agent Systems | Cloud Deployment

---

⭐ Judge Impact Statement

PaceBuddy demonstrates how multi-agent AI can move beyond conversations into real productivity execution.

It combines agent orchestration, cloud deployment, workflow automation, and tool integration into a practical system that solves everyday task management challenges.

This directly aligns with the competition goal of building API-based multi-agent systems that interact with multiple tools and structured data sources.