Case study
Feedlify: an AI feedback platform, built solo
A feedback platform that collects customer requests from everywhere they land, uses AI to merge the duplicates, and tells people when their request ships.
- Role
- Design and build, solo
- Status
- Launching 2026
- Built with
- React, Supabase, Clerk, OpenAI, Jira, Resend, Stripe, Sentry
The problem
Customer feedback arrives in many places at once. The same request turns up several times in different words, and the customer who asked for it rarely hears back when it finally ships.
What it does
Feedlify runs one loop: collect the feedback, triage it with AI, turn what's worth building into a Jira ticket, and tell the customer when it's done.
- Collects feedback from the places it arrives
- Uses AI to find requests that mean the same thing and merges them
- Creates Jira tickets for what's worth building
- When a ticket's status changes, notifies the customers who asked. The product manager sees the rendered email before anything is sent.
What I built
All of it, solo: the interface, the database with row-level security, sign-in, the Jira integration, background jobs, AI clustering, the notification engine, payments, and error tracking.
What it means for client work
The same pieces come up in small-business projects: AI that has to be right before anyone trusts it, integrations with tools a team already uses, and data that has to stay separate between customers. I'd rather learn those lessons on my own product than on yours.
Services
Let's find out whether this is worth doing.
Tell me what you need built, or what's wrong with the site you have. I'll reply by email, and if it's worth talking through, we'll set up a free 30-minute call.