Digital Twin AI

An AI-powered digital twin that mimics your communication style, knowledge, and personality. Built with FastAPI, Next.js, and AWS Bedrock.

What is this?

This application creates a conversational AI that represents you. It learns from your personal data (LinkedIn profile, facts about you, communication style) and responds to questions as if it were you. Think of it as your digital representative that can answer questions about your background, experience, and expertise.

How it works

The application consists of:

• Backend: FastAPI server that uses AWS Bedrock (Nova models) to generate responses based on your personal data
• Frontend: Next.js chat interface where users can interact with your digital twin
• Data Layer: Your personal information (LinkedIn PDF, facts, style guide) that personalizes the AI

When someone asks a question, the AI references your personal data to craft responses that sound like you and contain accurate information about you.

Architecture

The Digital Twin uses a modern architecture with multiple AI providers and flexible deployment options:

graph LR
    %% User Interface Layer
    subgraph "Frontend"
        UI[Chat Interface<br/>Next.js]
    end

    %% Infrastructure Layer (Cloud or Local)
    subgraph "Infrastructure"
        subgraph "Cloud (AWS)"
            CF[CloudFront CDN]
            APIGW[API Gateway]
            LAMBDA[AWS Lambda]
        end

        subgraph "Local Dev"
            LOCAL[FastAPI Server<br/>localhost:8000]
        end
    end

    %% Backend Application
    subgraph "FastAPI Backend"
        APP[Digital Twin API]
        CHAT[Chat Endpoint]
        %% RATE[Rate Limiter]
    end

    %% AI Services
    subgraph "AI Providers"
        BEDROCK[AWS Bedrock]
        OPENAI[OpenAI]
    end

    %% Data Storage (Flexible)
    subgraph "Data Storage"
        subgraph "Cloud Storage"
            S3D[S3 Personal Data]
            S3M[S3 Memory]
        end

        subgraph "Local Storage"
            FILES[Local Files<br/>backend/data/]
            MEMORY[Local Memory<br/>../history/]
        end
    end

    %% External Services
    subgraph "External APIs"
        AIAPIS[AI Provider APIs]
    end

    %% Connections - Cloud Path
    UI -.->|Cloud| CF
    CF --> APIGW
    APIGW --> LAMBDA
    LAMBDA --> APP

    %% Connections - Local Path
    UI -.->|Local| LOCAL
    LOCAL --> APP

    %% Common Backend Flow
    APP --> CHAT
    %% CHAT --> RATE
    CHAT --> BEDROCK
    CHAT --> OPENAI
    BEDROCK --> AIAPIS
    OPENAI --> AIAPIS

    %% Data Connections (Cloud)
    CHAT -.->|Cloud| S3D
    CHAT -.->|Cloud| S3M

    %% Data Connections (Local)
    CHAT -.->|Local| FILES
    CHAT -.->|Local| MEMORY

    %% Styling for dark backgrounds
    classDef frontend fill:#00bcd4,stroke:#ffffff,stroke-width:2px,color:#ffffff
    classDef aws fill:#ff6f00,stroke:#ffffff,stroke-width:2px,color:#ffffff
    classDef ai fill:#4caf50,stroke:#ffffff,stroke-width:2px,color:#ffffff
    classDef backend fill:#2196f3,stroke:#ffffff,stroke-width:2px,color:#ffffff
    classDef storage fill:#9c27b0,stroke:#ffffff,stroke-width:2px,color:#ffffff

    class UI frontend
    class CF,APIGW,LAMBDA,S3D,S3M aws
    class LOCAL,FILES,MEMORY storage
    class BEDROCK,OPENAI,AIAPIS ai
    class APP,CHAT backend

Loading

Key Features:

• Flexible deployment - Cloud (AWS serverless) or Local development
• Real-time responses from AI providers
• Multi-AI provider support (AWS Bedrock, OpenAI)
• Adaptive data storage - S3 for cloud, local files for development
• Security controls and validation

📖 For detailed architecture information and deployment guide, see Architecture & Deployment.

Data Management

Current Setup (Separate Private Repo)

This project is currently configured to keep personal data separate from the main codebase. Personal data is stored in a private repository and encrypted for deployment.

Alternative: Local Data Approach

A simpler local approach by committing personal data to backend/data directory:

1. Copy the templates from backend/data/personal_data_templates/ to backend/data/personal_data/

2. Edit the files with your actual information:

   • facts.json - Structured facts about you
   • summary.txt - Personal summary
   • style.txt - Your communication style
   • linkedin.pdf - Your LinkedIn profile as PDF
   • me.txt - Additional personal description (optional)

3. Copy prompt templates from backend/data/prompts_template/ to backend/data/prompts/

Quick setup command:

./scripts/setup-local-data.sh

📖 For detailed information about data files, formats, and the data loader system, see the Data Guide.

Customization

Avatar

To use your own avatar image:

1. Replace the avatar file in frontend/public/avatar.png with your own image
2. Recommended size: 200x200px or larger (square aspect ratio)
3. Supported formats: PNG, JPG, or any web-compatible image format

The avatar will automatically appear in the chat interface.

Personalize Names and Text

Edit frontend/app/page.tsx to customize:

• Digital twin name (default: "Luna")
• Your name and title
• Header and footer text

Getting Started - Local Development

⚠️ Cost Warning: Local development using OpenAI will incur API costs based on your usage. OpenAI charges per token (input/output). Monitor your usage at OpenAI's usage dashboard. Alternatively, you can use AWS Bedrock which also has associated costs.

Prerequisites

• Python 3.11+
• Node.js 18+
• OpenAI API key (or AWS credentials for Bedrock)

Setup Steps

1. Install dependencies:

# Backend
cd backend
uv sync

# Frontend
cd frontend
npm install

2. Set up your personal data (see Data Management section above)

3. Configure environment variables (optional):

For local development, you can use OpenAI instead of AWS Bedrock. Copy the example environment file and configure it:

cd backend
cp .env.example .env
# Edit .env and set AI_PROVIDER=openai and your OPENAI_API_KEY

See backend/.env.example for all available configuration options.

4. Start the backend (Terminal 1):

cd backend
uv run server.py

Backend runs on http://localhost:8000

5. Start the frontend (Terminal 2):

cd frontend
npm run dev

Frontend runs on http://localhost:3000

Test Your Local Setup

1. Open your browser to http://localhost:3000
2. You should see the chat interface with your avatar
3. Start chatting with your digital twin!

---

Cloud Deployment

⚠️ Cost Warning: Deploying to AWS will incur cloud infrastructure costs that are your responsibility. See the Cost Disclaimer section in the architecture guide for details.

Willing to deploy to AWS? See the Architecture & Deployment Guide for instructions. Once setup:

Quick deploy:

./scripts/deploy.sh dev

---

Project Structure

digital-twin/
├── backend/           # FastAPI backend
│   ├── app/          # Application code
│   ├── data/         # Personal data and prompts
│   └── tests/        # Backend tests
├── frontend/         # Next.js frontend
│   ├── app/          # Next.js app directory
│   └── components/   # React components
├── terraform/        # Infrastructure as Code
├── scripts/          # Deployment and utility scripts
└── docs/            # Documentation

---

License

See LICENSE file for details.

---

More screenshots and documentation coming soon...