Dram
Title
Dram, A Conversational Whisky Recommender That Starts With Taste
Summary
Dram turns everyday preferences—food, drinks, moods, and memories—into a practical whisky shortlist. It builds a structured flavour profile, explains why each bottle fits, and keeps model credentials safely on the server rather than exposing them in the browser.
Show Tech Stack
Tech Stack
- 🛠️ Frontend: Static HTML, CSS, and Vanilla JavaScript
- 🛠️ Backend: FastAPI, Pydantic, Uvicorn, and SQLite
- 🛠️ Recommendation: Structured taste profiles matched against a local whisky catalogue
- 🛠️ Models: Groq-hosted language models with an optional TypeSafe JEV engine
- 🛠️ Privacy: Provider credentials stay in server-side environment variables
- 🛠️ Data: WhiskeyProject flavour vectors with local conversation history