TrueFit

Inspiration

Buying a PC or baby gear takes hours of research. We wanted to help non-experts choose with clear reasons and review evidence, instead of relying on star ratings.

What it does

TrueFit turns your needs into a budget-checked shopping list, explains each choice, shows review patterns, and lets you adjust the plan before making the final decision.

How we built it

We built three agents with AWS Strands Agents SDK for gathering needs, editing results, and writing guides. Validated tools manage state; code handles budgets, ranking, and checks.

Challenges we ran into

We had to keep agent actions within category rules, prevent invented products, and explain review signals without claiming to detect fake reviews.

Accomplishments that we're proud of

We analyzed 43.9 million Amazon reviews, made recommendation decisions traceable, and reused the same agent code across PC building and baby care.

What we learned

Agents work best with narrow tools and explicit rules. Code should handle calculations and validation, while AI helps users express needs and understand results.

What's next for TrueFit

We plan to add live prices, expand product categories, improve catalog coverage, and recheck compatibility after swaps to help users make more informed purchases.