Orange Hyper is an RPG-style adaptive project-memory harness that starts as a lightweight intent compiler and grows project-specific memory, roles, tools, and verification loops only when repeated work proves they are useful.

Ollmo is a state-native runtime for local AI with Ollama, MLX, and llama.cpp. It makes supported model and media workflows durable and inspectable by recording which outputs exist, what remains open, and why.

Governed intent-to-execution compiler/runtime for L9 kernels, activation plans, execution contracts, execution graphs, validation evidence, and adapter-rendered AI workflows.