Neon Blueprints

Declarative Query Synthesis (DQS) is a library for intent-based data access where clients describe *what data they need* rather than *how to retrieve it*. The system consists of three core components: a pure-JSON DSL for expressing data requirements, an engine that compiles DSL to optimized SQL, and hooks for LLM-based DSL generation. Combined with Iris for response transformation, DQS enables end-to-end declarative data access from natural language intent to shaped response.

User interfaces have been artifacts—compiled, deployed, frozen in code. This paper proposes an architecture where interfaces are projections: runtime transformations defined in machine-readable schemas that language models can generate, orchestrators can observe, and systems can serve dynamically without deployment. The result is infrastructure where humans build the machine, and machines build the interface.

This paper introduces the Stratified Memory Architecture (SMA), a novel approach that decomposes machine memory into specialized layers—each with distinct storage structures, indexing strategies, and retrieval patterns.

The original Adam Whitepaper

The original whitepaper for project-iris