🐾 PetBloom 🌸

An AI-powered Pet Care Concierge that helps pet owners track health, remember important events, and receive personalized care recommendations.

📖 Overview

PetBloom is an AI-powered pet care platform built around an Agentic AI architecture combining long-term memory, tool calling, contextual retrieval, streaming LLM responses, and personalized recommendation generation.

Instead of simply storing pet information, PetBloom continuously observes daily care activities, remembers important events, analyzes historical patterns, and generates intelligent recommendations that help owners provide healthier and more consistent care.

The project demonstrates how modern AI agents can combine memory, reasoning, tool use, and contextual retrieval to solve real-world problems.

🚀 Live Demo

🌐 Live App: https://petbloom-40612.web.app

❓ Problem

Pet owners often struggle to maintain consistent care routines for their pets.

Important tasks such as vaccination schedules, medication, hydration, exercise, nutrition, and symptom tracking are usually managed through scattered notes, reminders, or memory.

Most existing pet apps act only as record-keeping tools and do not understand the pet's history or provide personalized guidance.

PetBloom addresses this by creating an AI-powered companion that remembers pet information, analyzes care patterns, and proactively helps owners make better decisions.

💡 Solution

PetBloom acts as an intelligent pet care assistant that combines AI reasoning, long-term memory, and personalized recommendations.

The system:

stores important pet information
learns from daily care activities
retrieves relevant history during conversations
provides personalized suggestions
helps owners maintain healthier routines

🤔 Why an AI Agent?

Traditional pet apps only store information.

PetBloom uses an AI Agent because pet care requires understanding context over time.

A normal chatbot answers based only on the current message.

PetBloom can:

🧠 Remember previous pet information
🔍 Retrieve relevant history
🛠️ Use tools to gather additional context
💭 Reason over multiple data sources
💡 Provide personalized recommendations

This allows PetBloom to move from passive tracking to proactive care assistance.

✨ Features

🤖 AI Concierge

AI-powered pet assistant
Context-aware conversations
Live Gemini streaming responses
Markdown support
Suggested follow-up questions
Rich AI response cards

🐶 Pet Management

Multiple pet profiles
Breed information
Medical history
Vaccination reminders
Lifestyle goals
Food inventory tracking

📊 Care Dashboard

Daily care checklist
Bloom Progress tracker
Quick activity logging
Upcoming reminders
Weather-aware insights
Recent activity timeline

🧠 Long-Term Memory

PetBloom doesn't simply answer questions.

It remembers:

allergies
symptoms
nutrition habits
exercise history
owner preferences
vaccination schedules

These memories are selectively retrieved to improve future AI responses.

📈 Weekly Analytics

Wellness score
Habit heatmap
Water trends
Exercise trends
Weight tracking
AI-generated care summaries

❤️ Health Tracking

Track:

meals
hydration
exercise
grooming
symptoms
medications
stool & urine logs

Everything contributes to personalized recommendations.

🎨 Modern UI

Responsive design
Tailwind CSS v4
Dark mode
Mobile navigation
Glassmorphism cards
Animated transitions
Loading animations
Accessible components

🤖 Agent Capabilities

PetBloom is designed as an AI Agent rather than a traditional chatbot.
It combines memory, tool usage, contextual retrieval, and reasoning to provide personalized pet care assistance.

🧠 1. Long-Term Memory

PetBloom maintains important information about each pet and retrieves relevant details when needed.

The agent remembers:

🐾 Pet profile information
- Name
- Breed
- Age
- Preferences
🏥 Health history
- Previous symptoms
- Medical observations
- Allergies
- Vaccination records
🍖 Care patterns
- Nutrition habits
- Exercise routines
- Owner preferences

This allows PetBloom to provide responses based on the pet's history instead of generic answers.

🛠️ 2. Tool Usage

PetBloom uses specialized tools to collect additional context and improve decision-making.

Available tools include:

🏥 Health Tracker

Tracks and retrieves:

Symptoms
Medications
Health observations
Medical history

🏃 Activity Tracker

Monitors:

Exercise patterns
Activity levels
Daily routines

🌦️ Weather Information

Uses environmental context to provide better recommendations.

Example:

"Since today's temperature is high, consider reducing outdoor activity and ensuring your pet has enough water."

Tools allow the agent to gather real-world information before generating recommendations.

🧩 3. Contextual Reasoning

PetBloom combines multiple sources of information before responding.

Instead of answering isolated questions, PetBloom understands patterns over time.

The agent analyzes:

Current User Query + Retrieved Memories + Pet History + Tool Outputs ↓ AI Reasoning ↓ Personalized Response

Example:

A decrease in activity alone may not be concerning.

However:

↓ Activity + ↓ Water Intake + ↓ Appetite + Previous Health History

may indicate that the owner should monitor the pet more carefully.

💡 4. Personalized Recommendations

After analyzing context, PetBloom generates actionable suggestions.

Recommendations can include:

🥗 Nutrition adjustments
💧 Hydration reminders
🏃 Activity suggestions
💊 Medication reminders
🩺 Health monitoring advice

The goal is to transform pet care from passive tracking into proactive assistance.

🏗️ AI Agent Architecture

                   User
                     │
                     ▼
              Chat Interface
                     │
                     ▼
             Agent Executor
                     │
     ┌───────────────┼────────────────┐
     ▼               ▼                ▼
 Tool Registry   Memory Search   Prompt Builder
     │               │                │
     └───────────────┼────────────────┘
                     ▼
             Gemini API / Mock AI
                     │
                     ▼
            Streaming AI Response
                     │
                     ▼
              Recommendation Engine
                     │
                     ▼
                 Dashboard

🧩 Agent Workflow

PetBloom follows an agentic workflow inspired by modern AI systems.

Observe
      ↓
Remember
      ↓
Retrieve Context
      ↓
Reason
      ↓
Generate Response
      ↓
Recommend Actions
      ↓
Update Memory

🛠️ Tech Stack

Technology	Purpose
React	Frontend
TypeScript	Type Safety
Vite	Build Tool
Tailwind CSS v4	Styling
Zustand	State Management
React Router	Routing
Gemini API	AI Responses
Canvas Confetti	Celebration Effects
Firebase	Deployment & Hosting

📂 Project Structure

src/
├── app/
├── assets/
├── components/
│   ├── layout/
│   └── ui/
├── config/
├── data/
├── features/
│   ├── auth/
│   ├── chat/
│   ├── dashboard/
│   ├── error/
│   ├── exercise/
│   ├── health/
│   ├── hygiene/
│   ├── landing/
│   ├── nutrition/
│   ├── onboarding/
│   ├── profile/
│   ├── reports/
│   └── settings/
├── layouts/
├── lib/
├── pages/
├── services/
│   ├── agent/
│   ├── api/
│   ├── mock/
│   ├── aiService.ts
│   ├── authService.ts
│   ├── geminiService.ts
│   ├── index.ts
│   ├── mapsService.ts
│   ├── petService.ts
│   ├── recommendationEngine.ts
│   └── weatherService.ts
├── store/
├── stores/
├── types/
├── utils/
├── App.css
├── App.tsx
├── index.css
├── main.tsx
└── vite-env.d.ts

🚀 Getting Started

Clone the repository

git clone https://github.com/riaannjoe/PetBloom.git

Move into the project

cd PetBloom

Install dependencies

npm install

Run locally

npm run dev

Build production version

npm run build

Preview production build

npm run preview

🔑 Environment Variables

Create a .env file.

VITE_GEMINI_API_KEY=YOUR_API_KEY
VITE_USE_MOCK_API=true
VITE_API_BASE_URL=http://localhost:3001/api

📸 Screenshots

🌟 Highlights

✅ AI Concierge

✅ Long-Term Memory

✅ Recommendation Engine

✅ Context Retrieval

✅ Live Gemini Streaming

✅ Weekly Analytics

✅ Multi-Pet Support

✅ Modern Responsive UI

🌐 Deployment

PetBloom is deployed using Firebase Hosting.

Production URL: https://petbloom-40612.web.app

The application is configured for public access so users and judges can interact with the AI assistant directly.

🔐 Security Considerations

PetBloom follows secure development practices:

🔑 API keys are stored using environment variables
🚫 Sensitive credentials are never committed to GitHub
🛡️ .env files are excluded using .gitignore
⚙️ Mock mode allows development without exposing production keys

🛤️ Development Journey

PetBloom started as a simple pet tracking idea.

During development, it evolved into an AI Agent system by incorporating:

Long-term memory
Context retrieval
Tool usage
Reasoning-based recommendations

The main learning was that useful AI systems are not created by simply connecting an LLM to an interface.

They require:

relevant context
memory
tools
clear responsibilities

🚧 Future Improvements

Voice conversations
Camera-based pet health detection
Smart collar integration
Wearable device support
Veterinary portal
Cloud synchronization
Push notifications
AI anomaly detection
Medication reminder automation

🏆 Kaggle AI Agents Capstone

This project was created as part of the Google × Kaggle AI Agents: Intensive Vibe Coding Capstone.

It demonstrates practical applications of:

AI Agents
Long-Term Memory
Tool Calling
Context Retrieval
Streaming LLM Responses
Recommendation Systems
Modern Frontend Architecture

YouTube demo video: https://youtu.be/dcVAso7lU_Q?si=p_yQFIUnSRirD9_p

👩‍💻 Author

Ria Ann Joe

GitHub: https://github.com/riaannjoe

📄 License

This project is licensed under the MIT License. See the LICENSE file for details.

Name		Name	Last commit message	Last commit date
.firebase		.firebase
docs		docs
public		public
src		src
.env.development		.env.development
.env.example		.env.example
.firebaserc		.firebaserc
.gitignore		.gitignore
.oxlintrc.json		.oxlintrc.json
LICENSE		LICENSE
README.md		README.md
firebase.json		firebase.json
index.html		index.html
package-lock.json		package-lock.json
package.json		package.json
tsconfig.app.json		tsconfig.app.json
tsconfig.json		tsconfig.json
tsconfig.node.json		tsconfig.node.json
vite.config.ts		vite.config.ts

Latest commit

History

Folders and files

Repository files navigation