# BenchLedger BenchLedger is a private maker inventory and project workspace. Its core tasks are keeping an accurate workshop inventory and planning and progressing a build. Projects and Inventory are peer destinations. Projects opens a visual library with Gallery and List views. Opening a project leads to Overview, Parts, Files and Build. Overview explains the project and its next action. Parts holds requirements, owned-stock matching and sourcing. Files groups revision-bound artifacts by purpose. Build presents preparation, assembly, verification notes and actual-stock-use review, with detailed planning tools available in context. A maker needs to distinguish what is owned, what is confirmed usable, what a project requires and what genuinely needs sourcing. Exact identity, compatibility, uncertainty and provenance remain available when a decision needs them. Recorded, ordered or delivered items must never be presented as physically counted or confirmed ready. Reservations hold availability; actual use is a separate reviewed change. Images identify projects without establishing build readiness. A selected image has an explicit role: design render, reference image or built-product photo. Optional assembly and PCB tools support detailed work. File previews and task progress do not certify physical checks. Shopping is a proposal, never purchase authority. The intended experience is professional, intuitive and easy to understand. Use simplified technical English, clear actions and progressive disclosure. The visual system uses warm white, graphite and restrained cobalt, with equivalent light, dark and system appearance settings. Avoid sales copy, decorative dashboards and gamification. The platform is React and Vite. Preserve accessible keyboard and touch use, responsive layouts, draft continuity, authentication and optimistic concurrency. UI changes share application behaviour with HTTP and MCP. Native mobile development is outside this web interface change. This public repository uses synthetic fixtures only. Private inventory, hosts, credentials and runtime data stay outside source control. Implementation, verification and deployment are separate outcomes.