apps/www/_customers/bree.mdx
Bree is a Canadian FinTech startup focused on helping over 20 million Canadians living paycheck to paycheck access better, faster financial products. As the company scaled, it needed to move quickly, experimenting, shipping features, and iterating on user feedback in near real time. But their initial backend stack was holding them back.
Bree had originally built their infrastructure on Fauna, a NoSQL database with a proprietary query language (FQL). The limitations quickly surfaced.
<Quote img="ryan-han-bree.png" caption="Ryan Han, Head of Engineering and AI, Bree"> Fauna’s custom language, FQL, introduced friction in our development process. It took time to learn and made it harder for our team to iterate and ship at the speed we needed. </Quote>Engineering velocity suffered. Teams couldn’t iterate on product features without hitting roadblocks, and performance tuning became an ongoing challenge.
And when Fauna announced its end of life, it only added urgency. Bree needed a replacement quickly, and it had to be developer-friendly, scalable, and suitable for AI-powered product development.
Bree had long been aware of Supabase. Supabase had earned its reputation among builders and indie developers for its ease of use and instant setup. But what sold Bree was how well Supabase aligned with their long-term technical goals.
<Quote img="ryan-han-bree.png" caption="Ryan Han, Head of Engineering and AI, Bree"> Builders around the world use Supabase for exciting projects. It is so intuitive that developers can start building immediately and ship production-ready features in a day. </Quote>After a thorough evaluation of vendors, Bree chose Supabase for several reasons:
Performance benchmarks and onboarding experience also played a key role. “Supabase had the best onboarding experience, the best performance benchmarks, and the best developer experience,” said Han.
Even beyond the product, Bree valued Supabase’s commitment to supporting startups. “Their attitude toward startups like us made the whole package very attractive.”
Migrating from Fauna’s NoSQL architecture to a relational Postgres database required careful coordination. Bree took a phased approach, designed to minimize disruption, avoid downtime, and ensure full data integrity along the way.
Phase one focused on replicating Fauna’s schemas in Supabase using Postgres jsonb columns to accommodate Fauna’s document-style structure. Bree built ETL pipelines and verification scripts to ensure data consistency across both platforms. This dual-write phase ensured every read and write operation was fully synchronized.
One of the primary challenges was handling type mismatches. Bree built conversion scripts to synchronize NoSQL data types with Postgres.
Once they verified the sync was intact, Bree were able to switch off Fauna and have Supabase as the main primary database.
Phase two, now underway, is focused on modularizing the data model, splitting large documents into normalized tables. This will allow Bree to take full advantage of Postgres performance and query flexibility.
The results of Bree’s migration to Supabase have been immediate and measurable. Engineers can ship faster, performance is up across the board, and Supabase has become a foundational part of the team’s AI-driven development workflow.
<Quote img="ryan-han-bree.png" caption="Ryan Han, Head of Engineering and AI, Bree"> After migrating to Supabase, we saw immediate lift in developer velocity. Engineers can iterate on features much faster by writing Postgres SQL queries. </Quote>Latency improvements are notable. Bree has observed double-digit reductions in response times across core functions and expects further gains as their architecture is optimized for Supabase.
<Quote img="ryan-han-bree.png" caption="Ryan Han, Head of Engineering and AI, Bree"> It already has double-digit latency improvements. With further optimization, some functions were 10x faster. </Quote>Supabase has also enhanced Bree’s ability to develop AI-first features. They’re able to use the Supabase MCP Server to power their development in Cursor. As Bree expands its AI initiatives, Supabase’s native support for vector databases and seamless integrations make it the clear choice for long-term scalability.
Bree’s journey reflects the reasons developers and fast-moving teams choose Supabase:
pgvector out of the box, making it ideal for AI-powered developmentStart your journey today at www.supabase.com