LocalStar
AI-powered review and reputation management platform designed to help businesses collect reviews, route private feedback and monitor reputation.
- Year
- 2025
Stack
- Next.js
- Node.js
- AI
- Supabase
- APIs
01
Overview
LocalStar helps local businesses grow their public reputation while catching problems in private, before they become one-star reviews.
02
Problem
Small businesses only hear from customers at the extremes, and usually too late. Asking for reviews manually does not scale, and one bad review outweighs ten silent happy customers.
03
Idea
One collection link that behaves differently depending on sentiment: delight becomes a public review, friction becomes a private ticket the owner can actually fix.
04
Architecture
Next.js frontend, Node.js services and Supabase Postgres with row-level security per business. Review sources sync through scheduled jobs; sentiment classification runs inline on submission.
05
Design
The customer-facing flow is two taps and no login. The owner dashboard is dense but calm — a single reputation line, an inbox, and one action per item.
06
Development
Multi-tenancy was designed first: every table is scoped to an organisation, and all access goes through policies rather than application-level filtering.
07
AI / Technical Challenges
Sentiment routing has to be conservative — misrouting an angry customer to a public review is unrecoverable. Thresholds are tuned to prefer private routing when confidence is low.
08
Results
A repeatable collection loop owners can run without thinking about it.
09
What I learned
In tools for non-technical operators, defaults are the product. Every setting you add is a support ticket.
10
Next steps
Deeper platform integrations and competitor benchmarking per location.