AI Roaster 😈

AI Roaster is an AI-powered roasting platform that generates witty and hilarious "roasts" for team members based on their unique characteristics. This project combines a serverless backend with dynamic AI processing, voice-based interactions, and customizable roast outputs.

How We Roasted Our Teammates 😂

Features

• 🌶 Funny and Personalized Roasts: Create tailored and hilarious roasts for team members using their characteristics and data.
• 🔥 Dynamic Roast Image Generation: Generate roast-themed images unique to each team member.
• 🌐 Web Application: A polished web interface for users to interact with the roasting experience.
• 🔗 Unique URLs for Roasts: Share roasts easily using one-click URLs.
• 🎙 Voice Output Support: Listen to your roast via Text-to-Speech (TTS) audio.

And it’s totally SERVERLESS! 😎

Tech Stack

• Backend: Python (Flask), AWS Lambda
• Frontend: HTML, CSS, JavaScript
• Serverless Framework: AWS infrastructure setup and deployment
• Dependencies:
  • Flask, Werkzeug, OpenAI API
  • Boto3 for AWS interactions

Getting Started

Prerequisites

1. You will need the following tools:

• Python 3.10 or later
• Node.js for Serverless Framework integration
• AWS CLI configured with the required permissions

2. Create a Bucket to store assets:

   • Create an S3 bucket in your AWS account to store the roast images and other assets.
   • Update the ASSET_BUCKET value in the .env file with the bucket name.

3. Create a Deployment Role:
   The GitHub workflow uses GitHub AWS OIDC.

   • Create an IAM role in your AWS account for GitHub Actions.
   • Update the GITHUB_CI_ROLE value in the .env file (GitHub Actions will use this role to deploy the project).

4. Set Up the OpenAI API Key:

   • Generate an OpenAI API Key.
   • Store the key in AWS Secrets Manager.
   • Update the OPENAI_KEY_SECRET_NAME in the .env file with the key's secret name.

5. Set Other Environment Variables:

   • Update ASSET_BUCKET, AWS_ACCOUNT and REGION in the .env file.

Training data

To bring the roasting experience to life, you need to provide descriptions and images for your teammates. Follow the steps below to populate the database:

1. Create Roast Descriptions:

   • Provide a detailed description of the person, including quirky traits, unique experiences, and anything that can serve as roast material—bonus points if you include pre-written roasts!
   • Save each roast in a separate .txt file within the src/roasts directory.
   • Use clear, descriptive file names (e.g., john_doe.txt).

2. Add Corresponding Images:

   • For each teammate, provide their image.
   • Save the images in the src/static/images/team directory.
   • Use the same naming convention as the .txt files (e.g., john_doe.png or john_doe.jpg).

Running Locally

1. Set Up the Environment:

   pip install poetry
   poetry install

2. Configure environment variables: Ensure your .env file is set up correctly and has the correct values.
3. Start the Server:

   python -m src.main

4. Open the application in your browser at:

   http://127.0.0.1:5000/
   

Deployment

The project uses the Serverless Framework for deployment on AWS Lambda, with GitHub Actions for deploying to AWS. To deploy:

• Commit your changes to the master branch.
• The GitHub Actions pipeline will deploy to AWS and host your website.

Project Structure

Ai-Roaster/
├── src/
│   ├── main.py                # Flask app and API handlers
│   ├── bedrock.py             # Core AI logic for finding team members and roasting
│   ├── roasts/                # Roast content files
│   ├── static/                # Frontend assets (CSS, JS, images)
│   ├── templates/             # HTML templates for the web app
├── serverless.yml             # Serverless configuration
├── pyproject.toml             # Python dependencies (Poetry)
├── package.json               # Node.js dependencies

Contributing

1. Fork the repository and create a branch for your feature.
2. Submit a pull request with detailed changes and test cases.

License

This project is licensed under the MIT License.