Roaster is a backend-driven AI application that analyzes user data from Spotify, YouTube, and Chess.com to generate contextual humorous insights using a locally hosted LLM.
The system integrates multiple third-party APIs and performs structured data processing before generating AI responses.
- Node.js (Backend Service)
- Ollama (Local LLM Inference)
- REST APIs (Spotify, YouTube, Chess.com)
- JSON Data Processing
Client → Node.js Backend → API Data Fetching → Data Processing → Ollama LLM → Generated Response
- Multi-source API integration
- Structured JSON data transformation
- Prompt-engineered AI output
- Local LLM deployment (no external AI API dependency)
- Install Node.js
- Install Ollama and download required model
- Configure environment variables
- Run backend server
- Send request to generate roast
- Rate limiting & authentication
- Response caching
- Model optimization
Run npm i to install the dependencies.
Run npm run dev to start the development server.
