Show HN: Yana – A local-first, human-governed AI runtime and IDE in Rust (github.com)

Yana is a groundbreaking AI runtime and integrated development environment (IDE) developed by Vũ Văn Tâm from Vietnam. Positioned as a control plane between AI models and their operational capabilities, Yana ensures that any action proposed by an AI is subject to human governance and strict permissioning protocols. Through a structured pathway, Yana allows AI models like Claude, GPT, and Gemini to express intentions while enforcing rules that mandate human approval before execution. This framework not only limits AI autonomy but also establishes robust oversight through features like real-time telemetry, a hash-chained audit log, and an independent monitoring system called Giám Thị that can halt actions at any point. Significant for the AI/ML community, Yana promotes a "local-first" architecture, emphasizing user control and security in AI interactions. Built in Rust, the runtime supports a myriad of tasks with 40 subcommands and incorporates a wide array of tools for project management, such as a desktop workspace, terminals, version control with Git, and a chat interface—all operating under the same governance framework. With over 2,000 skills designed to train AI agents, Yana not only aims to bolster the capabilities of AI systems but also insists on a more responsible approach to their deployment in real-world applications.