We build the data & AI infrastructure other companies depend on.

Our story

Arcade Data was founded by engineers who had spent more than a decade at the centre of the multi-model database movement. They believed the industry had been pushed into an awkward compromise: either pick a single specialised store and accept a fragmented stack (one database for documents, another for graphs, another for vectors), or pay the operational cost of running them all.

That conviction became ArcadeDB — a single, ACID-compliant database engine where graph, document, key-value, full-text, vector and time-series workloads share the same storage. Released under Apache 2.0 with a deliberate, written commitment never to relicense or add enterprise-only features, ArcadeDB has grown into one of the most actively developed open-source multi-model databases in production today.

In 2024 we extended that thesis upwards. As application teams began assembling fragile stacks of point services to build LLM-powered products — one tool for workflows, another for vector search, another for auth — we shipped ArcadeBrain: an AI-native backend platform that consolidates database, workflows, LLM orchestration, authentication and billing into a single deployable system. It is built on the same engine as ArcadeDB, and shares its open philosophy.

Today we operate as a small, senior team of engineers, researchers and operators — distributed across more than thirty countries, anchored in our London office. We measure success the way our customers do: in latency, in correctness, in the systems that stay up when it matters.