Building Library Ground / Aerie: A Windows-Native Offline Ledger Engine by SAGEWORKS AI

SAGEWORKS AI has developed a software system called Library Ground / Aerie, specifically designed to run natively on Windows machines without requiring additional software such as Rust, Cargo, Python, Node.js, or an internet connection. The Windows-native edition is tailored to preserve knowledge, structure it, question it, audit it, and extend it locally, without relying on a cloud model.

The system consists of three cooperative roles: Herald, Meridian, and Forge. Herald serves as the front communication layer, receiving commands, answering questions, recording conversations, and displaying evidence. Meridian governs writes, checks references, maintains ordering, and ensures the integrity of the store. Forge constructs views from preserved sources, proposes labels, identifies gaps, and creates specialized workers.

The Windows-native edition is built around a command-line knowledge system that functions using three logical roles running within a single local PowerShell process. These roles are recorded through an internal event timeline called AERIE, making the system lightweight and easily inspectable. Upon closing the window, no background service remains active.

The launcher for Library Ground / Aerie includes several command files, such as START_LIBRARY_GROUND.cmd, CHECK_SYSTEM.cmd, windows-native/LibraryGround.ps1, windows-native/Test-LibraryGround.ps1, START_HERE.md, WINDOWS_NATIVE_GUIDE.md, and NATIVE_VALIDATION.md. These files facilitate launching the system, checking the local Windows environment, running the active engine, conducting native self-tests, providing a first-session guide, operating and troubleshooting guide, validation notes, and more.

The system's evidence-based approach is built around the principle that if it does not have evidence, it should ask for evidence instead of inventing an answer. When creating a subject and adding information, it is stored as a source. Later, when asking a question, Herald searches the selected subject and returns matching source lines with references, making the answer traceable. This approach ensures explainability, auditability, and offline usability.

A typical first session with Library Ground / Aerie might involve creating a subject, pasting information, checking the status, building views, identifying gaps, and teaching the system about specific concepts. The /audit command verifies ordering, checksums, source presence, and references, ensuring the system's integrity. However, it is essential to understand that integrity is different from truth, as checksums prove the text was not accidentally changed but cannot confirm the original claim's factual correctness.