AIPOCH Open-Science 1.0 Ships Portable Research Context

The compelling idea here is treating research provenance as a shareable artifact, not scattered chat logs and scripts. It could make AI-assisted science more auditable, though `.science` is a structured handoff record—not a guarantee of reproducibility.

• –Packages preserve research context across computers while keeping imported sessions read-only and credentials isolated.
• –Python/R execution, literature tools, scientific connectors, and multiple model providers make this broader than an AI writing assistant.
• –Apache-2.0 licensing and local-first storage give institutions more control over deployment and data flow.
• –Reproduction still depends on complete inputs, dependencies, runtimes, and external data; the receiving machine does not automatically rerun the work.
• –The current product is stronger for execution and provenance than real-time team collaboration.