SynapseAI: Orby AI

AI Buddy: Your Proactive Personal Companion Against Loneliness

Tagline
AI Buddy is a multi-agent personal companion designed to proactively combat loneliness by providing personalized assistance, support, and engagement through intelligent scheduling, task management, well-being monitoring, and empathetic conversation.

Description & Why it Matters
In an increasingly connected yet isolated world, many struggle with loneliness. AI Buddy is an innovative solution that leverages generative AI to create a proactive, empathetic digital companion. It goes beyond simple chatbots by integrating with a user’s digital life—managing tasks, fostering mental well-being, and providing personalized interaction. The goal is to reduce feelings of isolation while enhancing productivity and focus.

Technical Excellence

End-to-end demo with a working web interface, containerized backend services, and deployed agents.

Multi-agent architecture, each service encapsulated in its own Docker container.

Live chat with real-time Gemini responses, calendar integration, task management, and well-being monitoring.

Solution Architecture & Documentation

Repository structure is modular and clear:

your-ai-buddy-hackathon-repo/
├── ai-buddy-platform-agent-orchestrator/ # Core orchestration logic
├── ai-buddy-platform-grounding-service/ # Context grounding
├── ai-buddy-platform-calendar-manager/ # Calendar sync
├── ai-buddy-platform-outlook-mail-checker/ # Mail integration
├── ai-buddy-platform-assignment-tracker/ # Task tracking
├── ai-buddy-platform-mental-health-checker/ # Check-ins
├── ai-buddy-platform-human-approval-workflow/ # Human-in-the-loop
├── ai-buddy-platform-memory-ingestion-service/ # Memory embedding
└── README.md # Setup, architecture diagram

README includes setup instructions, build and run commands, and architecture diagrams for clarity.

Gemini Integration

Gemini 1.5 Pro for natural conversation and reasoning.

Multimodal support: text, task context, and calendar data.

Tool calling for scheduling, mail checking, and memory services.

Chaining across agents with orchestrator managing requests.

Human-in-the-loop approval agent for sensitive actions.

Impact & Innovation

Reduces loneliness by offering empathetic, personalized conversation and support.

Promotes productivity with proactive scheduling, reminders, and task tracking.

Encourages mental well-being through check-ins and exercises.

Innovative integration of wellness and productivity in a multi-agent framework.

Designed for extensibility with Vertex AI, ADK, and Agentspace runtime.

Features

Conversational AI powered by Gemini.

Calendar and email integration.

Mental health check-ins and well-being monitoring.

Task and assignment tracking with progress visualization.

Human approval workflow for user trust and control.

Limitations

Voice transcription simulated; no live speech-to-text yet.

Persistence limited to session-level storage.

Prototype UI, not fully polished for production.

Deployment Status

Services containerized with Docker.

Backend deployed on Google Cloud Run.

Cloud SQL for structured data storage.

Architecture ready to extend with Vertex AI Agentspace orchestration.

Tech Stack

Languages/Frameworks: Python, Flask

AI: Gemini API, Vertex AI, google-generativeai

Infrastructure: Docker, Cloud Run, Nginx, Cloud SQL

Libraries: flask, flask-cors, google-cloud-aiplatform, pg8000

your-ai-buddy-hackathon-repo/
├── ai-buddy-platform-agent-orchestrator/
│ ├── main.py # The core orchestrator logic (from pseudocode)
│ ├── requirements.txt # Python dependencies (flask, google-cloud-aiplatform, etc.)
│ ├── Dockerfile # Instructions to build the Docker image
│ └── .env # Environment variables for local development (NOT for deployment)
├── ai-buddy-platform-grounding-service/
│ ├── main.py # Grounding logic
│ ├── requirements.txt
│ ├── Dockerfile
│ └── .env
├── ai-buddy-platform-calendar-manager/
│ ├── main.py # Calendar management logic
│ ├── requirements.txt
│ ├── Dockerfile
│ └── .env
├── ai-buddy-platform-outlook-mail-checker/
│ ├── main.py # Outlook mail checking logic
│ ├── requirements.txt
│ ├── Dockerfile
│ └── .env
├── ai-buddy-platform-assignment-tracker/
│ ├── main.py # Assignment tracking logic
│ ├── requirements.txt
│ ├── Dockerfile
│ └── .env
├── ai-buddy-platform-mental-health-checker/
│ ├── main.py # Mental health check-in logic
│ ├── requirements.txt
│ ├── Dockerfile
│ └── .env
├── ai-buddy-platform-human-approval-workflow/
│ ├── main.py # Human-in-the-loop logic
│ ├── requirements.txt
│ ├── Dockerfile
│ └── .env
├── ai-buddy-platform-memory-ingestion-service/
│ ├── main.py # Memory embedding and ingestion logic
│ ├── requirements.txt
│ ├── Dockerfile
│ └── .env
└── README.md # Overall project readme