Solstice should read as a platform for building, running, and surfacing intelligence systems. The core story is simple: we generate usable operating environments, we build replayable runtime memory, and we turn both into interfaces people can actually use.
Structured, scenario-rich data packs for demos, analytics, simulations, copilots, and workflow products where real operational data is restricted, expensive, or unusable.
DeltaStore and TSC make state inspectable, replayable, compressible, and branchable, so long-running AI systems can persist memory without turning into opaque black boxes.
Iris is where the stack becomes usable: conversation, voice, vision, and interaction tied to memory, tools, and product context instead of isolated one-shot chat.
Explore the public artifacts that show how Solstice works in practice: dataset releases, runtime demos, and technical research tied to the same platform.
See public dataset packs, evaluate structure and realism, and use that as the cleanest entry point into custom work.
Go straight to DeltaStore and the TSC papers if you need to understand the memory and compression architecture.
Use the site as a clearer front door for partnerships, pilots, and scoped engagements instead of a gallery of experiments.