Most organizations operate across fragmented conversations, decisions, meetings, and updates. Relay is an operational system designed to help teams understand what is happening, coordinate execution, and adapt as conditions change. The first step is reconstructing operational reality from the signals an organization creates every day.
Relay was built using an AI-native development workflow centered on Cursor and Cursor Cloud Agent. The codebase spans roughly 20-30k lines of TypeScript across signal processing, memory persistence, entity resolution, hybrid grouping, feed synthesis, a Supabase persistence layer, validation suites, and a deployed Next.js application. Development followed a primitives-first approach where Signals, Memory, Entity Resolution, Hybrid Grouping, and the Reality Feed were designed and validated independently before being integrated into a working system.
Today, Relay's integrations are intentionally simple. Slack messages are mapped into signals using lightweight rules and metadata, then grouped into narratives using entity references, memory, and hybrid grouping. The value is not the keyword detection itself. The value is that Relay can already observe events, remember context, connect related information, and explain how a narrative was formed.
Over time, additional signal sources such as Gmail, Calendar, QuickBooks, Rippling, CRM systems, and meeting transcripts can contribute to the same operational graph. Relay can then connect people, money, customers, communication, and work into shared narratives, helping organizations move from observation to understanding, coordination, and adaptation.