MoodSpot-AI

Analyze drawings to reveal your mood and receive wellness advice.

Comment

MoodSpot AI:

Understand your emotions when words are not enough. MoodSpot AI is an innovative web application that uses artificial intelligence to analyze your drawings, revealing your underlying mood and providing personalized wellness advice. In a world where we are often disconnected from our true feelings, MoodSpot serves as a creative tool for self-discovery. It provides a safe and private space to express emotions non-verbally, helping users become more aware of their mental state and offering gentle, supportive guidance.

Key Features:

  • Creative Emotional Expression: A simple and intuitive digital canvas to draw whatever you're feeling.
  • Instant AI Mood Analysis: Leverages OpenAI's powerful vision models to provide real-time analysis of your drawings.
  • Personalized Wellness Recommendations: Receive actionable, supportive advice tailored to your detected mood.
  • Historical Mood Journey: Every analysis is securely stored on TiDB Cloud, allowing you to track your emotional patterns and journey over time.

Tech Stack & Services: MoodSpot AI is built on a modern, scalable, and AI-native tech stack...

  • Frontend: React
  • Database: TiDB Cloud (Serverless)
  • Development Assistant: Built with the AI-powered capabilities of Kiro
  • AI Mood Analysis: OpenAI API (GPT-4 Vision)

How MoodSpot AI Works: An Innovative AI Workflow MoodSpot AI is more than just an app- it's a complete, multi-step agentic process that transforms a simple drawing into actionable insight. This workflow showcases a powerful integration of modern AI and data technologies.

The Multi-Step Flow: Our application follows an automated, end-to-end process that aligns perfectly with the architecture of modern AI agent

  1. Ingest & Index Data (The Drawing): The process begins when a user creates a drawing. This image—a rich, unstructured piece of data—is ingested by the application. This visual data serves as the primary input for our AI agent.
  2. Invoke External Tools (OpenAI API Call): The application invokes an external tool: the OpenAI Vision API. The drawing is sent to the API, which acts as a specialized "tool" for emotional analysis, interpreting the colors, lines, and shapes to derive meaning.
  3. Chain LLM Calls (Analysis & Recommendation): The OpenAI service performs its analysis and returns a structured JSON object containing the primaryMood, confidence, and a descriptive analysis. Based on this result, a subsequent process generates personalized wellness recommendations, creating a chain of intelligent actions.
  4. Search Your Data (TiDB Cloud Integration): The structured analysis result is then indexed and stored in TiDB Cloud database. This crucial step transforms the ephemeral analysis into persistent, queryable data. TiDB's robust indexing allows for efficient retrieval, making it possible to query past moods and build a historical emotional timeline for the user.
  5. Build a Multi-Step Flow (The Complete User Experience): This entire sequence is wired together into a single, seamless user experience. From the initial brushstroke to the final display of personalized advice and the saving of the mood entry, the process is a fully automated, multi-step flow. The user simply draws and receives insight, while the underlying agent handles the complex chain of ingestion, tool invocation, data storage, and presentation.

Data Flow Summary: [User Draws on Canvas] --> [Image Ingestion] --> [OpenAI API Call] --> [AI Mood Analysis] --> [TiDB Cloud Storage] --> [Display Results & Recommendations]

Leveraging TiDB Cloud for a Scalable Data Backbone: TiDB Cloud is the core of MoodSpot AI's data persistence and analytics capabilities.

  • Scalable Storage: TiDB serverless cluster, which automatically scales to handle user growth without any manual intervention, is used. This is perfect for an application where usage can be unpredictable.
  • Data Reliability: Every mood analysis is stored in a 'mood_entries' table. TiDB's distributed architecture and ACID compliance ensure that this data is always consistent and available.
  • Foundation for Future Features: Storing mood data in TiDB allows for powerful future features, such as long-term trend analysis, pattern recognition, and even vector search to find drawings with similar emotional characteristics.

Getting Started & Running the Project: This project was bootstrapped with Create React App and requires Node.js to run. The core logic relies on API keys for OpenAI and TiDB Cloud, which should be stored in a .env file.

A brief overview of the local setup process:

  1. Clone the repository
  2. Install dependencies using npm install
  3. Configure API keys in a local .env file
  4. Run the development server with npm run dev

Backend Setup and Database

Starting the Backend Server

To run the MoodSpot AI backend server:

# Navigate to the backend directory
cd backend

# Install dependencies (if not already installed)
npm install

# Start the development server
npm run dev

Make sure your .env file contains your TiDB Cloud connection credentials.

TiDB Database Schema

Table Structure for mood_entries:

CREATE TABLE mood_entries (
  id BIGINT PRIMARY KEY AUTO_INCREMENT,
  user_id VARCHAR(255) NOT NULL,
  mood VARCHAR(50) NOT NULL,
  confidence DECIMAL(3,2),
  analysis TEXT,
  recommendations TEXT,
  created_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP
);

Sample INSERT Query:

INSERT INTO mood_entries (user_id, mood, confidence, analysis, recommendations) 
VALUES 
('user123', 'Happy', 0.92, 'Drawing shows bright colors and upward strokes', 'Continue creative activities');

Sample SELECT Query:

SELECT mood, confidence, created_at 
FROM mood_entries 
WHERE user_id = 'user123' 
ORDER BY created_at DESC 
LIMIT 5;

Database Integration

MoodSpot AI uses TiDB Cloud Serverless for scalable, reliable data storage. Every mood analysis is automatically saved with:

  • User session tracking for personalized mood history
  • Confidence scores from AI analysis (0.0 to 1.0 scale)
  • Detailed analysis text explaining the mood detection
  • Personalized recommendations based on detected emotions
  • Timestamps for tracking emotional patterns over time

The backend automatically handles database connections, error handling, and data validation to ensure reliable mood data persistence.

Due to the requirement for private API keys, a live, publicly accessible demo is not provided. Please refer to the demo video for a complete walkthrough of the application's functionality.

Built With

Share this project:

Updates