Inspiration

To simplify tech shopping by matching real user needs with the ideal laptop specs—no more guesswork or endless comparison charts.

What It Does

Parses your requirements (e.g. “AutoCAD 3D modeling under $1,500”), fetches recommended feature thresholds, and delivers the best-fit laptop plus a clear two-line rationale.

How We Built It

Backend: Flask API

Data: DynamoDB for need→feature mappings, CSV catalog for laptop specs

Frontend: Plasmo + React + Mantine

AI: Gemini LLM for natural-language parsing and concise explanations

Challenges We Ran Into

Ensuring robust numeric & categorical merging across varied feature formats

Handling edge cases when user phrasing didn’t map cleanly to known categories

Accomplishments We’re Proud Of

Smooth end-to-end flow from free-text input to actionable recommendation

Two-line, confidence-toned explanations that users consistently rate as clear and helpful

What We Learned

LLM parsing can be surprisingly reliable—but always needs fallback checks for malformed JSON

DynamoDB batch operations scale well for our small-to-medium rule set

What’s Next

Add support for user reviews & real-time price APIs

Expand to cover other categories (gaming PCs, desktops, workstations)

Build a lightweight web UI for drag-and-drop want selection

Built With

Share this project:

Updates