Prisma

Customer story

How Reflag ships reliably with a lean team on Prisma ORM

Reflag, a feature-management platform out of Copenhagen, runs its product on Node.js and PostgreSQL with a team of eight. It uses Prisma ORM to manage the complex relationships between features, users, and feedback without slowing development down.

For a small team, the data layer can't become the work. Prisma keeps queries type-safe and migrations schema-driven, so Reflag keeps shipping features reliably as the product grows.

PrismaCustomer
Reflag
Node.jsPostgreSQLGoogle Cloud

About Reflag

Reflag is a feature-management platform for product engineering teams, bringing feature flags, user feedback, and adoption metrics together in one place. What started three years ago in Copenhagen is now a live product run by a team of eight, built on Node.js services talking to PostgreSQL on Google Cloud.

  • Industry

    Developer tools / feature management

  • Team size

    8

  • Stack

    Node.js, PostgreSQL, Google Cloud

  • Using

    Prisma ORM

The challenge

Reflag's product lives on complex relationships between features, users, and the feedback tied to them. Modeling and querying that data across their Node.js services meant holding two things together that usually pull apart: moving fast, and staying type-safe as the schema kept growing.

With a team of eight, there's no room for the data layer to slow things down:

A data model with real depth. Features, users, and feedback all relate to each other, and every new capability adds relationships to keep consistent.

Velocity that can't drop. A small team ships often, so queries and migrations have to stay easy to change.

Type safety that has to hold. Mistakes caught at compile time are mistakes that never reach a customer.

The database is central to the product. It just can't be the thing that slows the product down.

Why Reflag chose Prisma ORM

Reflag reached for Prisma ORM to make working with PostgreSQL fast and safe. The draw wasn't a single feature, it was how the schema, the generated client, and relational queries work together to keep a small team productive.

Type-safe from the schema

A generated, fully typed client catches mistakes before they ship, no manual setup to keep in sync.

Relational queries made simple

Prisma's include and select turn complex joins into queries that are easy to write and easy to read.

One schema the team shares

The schema documents the data model in one place, so eight people stay aligned on how the data fits together.

“We rely on Prisma heavily for relational queries. It's so intuitive and makes complex queries easy to manage, allowing us to build advanced data retrieval patterns with minimal effort.”

Ron Cohen

Co-founder & CTO, Reflag

Reflag

How they use Prisma

Reflag uses Prisma ORM as the data-access layer across its Node.js services. As the product evolves, the team uses schema-based migrations to experiment with the database structure, then leans on Prisma's include and select to handle the relational queries the product depends on, all through a generated type-safe client.

The shared schema doubles as documentation. Instead of tribal knowledge about how features, users, and feedback connect, there's one model everyone works from, which keeps a small team moving in the same direction.

The results

With Prisma handling the data layer, Reflag ships reliably at a size where every hour counts.

8-person team

Shipping features reliably without a dedicated database team.

Complex queries, simplified

Relational joins handled through include and select.

One shared schema

The data model documented in one place the whole team works from.

Prisma makes database management incredibly easy. For example, when we needed to implement OAuth, which involved creating multiple new database tables, we just defined them in our Prisma schema, ran a migration, and everything was ready. Prisma's built-in type safety helps us avoid mistakes that happen with manual setups.
Reflag

Ron Cohen, Co-founder & CTO at Reflag

Build on the same stack as Reflag

If an eight-person team can manage complex relational data and keep shipping, yours can too.

  • Free to start, no credit card required
  • One TypeScript stack your agent can operate end to end
  • Trusted by 500K+ developers