Fraud detection dashboard, from first report to block
A web dashboard for the people who handle fraud: log a report, open an investigation, and put confirmed bad actors on a blocklist.
- Year
- 2025
- Type
- Web app
- Built with
- Next.js, Prisma, MongoDB, Clerk, Svix webhooks
What it does
Fraud work is a pipeline. Someone reports a suspicious entity, an analyst investigates, and if the report holds up the entity is blocked. The dashboard gives each stage its own record so nothing falls between them.
Data model
| Record | What it stores |
|---|---|
| Report | The reported entity, its type, the details, and a status. |
| Investigation | The entity, the investigator, a status, and start and end dates. |
| Blocked entity | The entity, who blocked it, and when. |
How it is built
- Next.js App Router for the dashboard and its API routes.
- Prisma on MongoDB for typed database access.
- Clerk for sign-in, with Svix-verified webhooks that keep the local user table in sync with Clerk.