> ## Documentation Index > Fetch the complete documentation index at: https://docs.atconseil.info/llms.txt > Use this file to discover all available pages before exploring further. # Organize at scale > Seed the folder tree automatically, file cases in bulk, read a flat view with full paths, and order cases by hand — built for projects with thousands of test cases. Filing cases one by one is fine for a demo, not for a real project with hundreds or thousands of test cases. This page covers everything that makes the Repository usable at scale — all of it **additive and reversible**, none of it modifying Azure DevOps. ## Auto-seed the folder tree Rather than build the tree by hand, **seed it** from a structure you already have: * **From Area Paths** — each area path segment becomes a folder (trimmed, empty segments ignored). * **From the plan/suite hierarchy** — your existing plans and suites become folders. A case that appears in several suites is filed once, in the first one encountered (the *single folder* rule). Seeding is **strictly additive**: existing folders are reused by name, never renamed, moved or deleted, and a case already filed is never moved. It is **idempotent** — running it again on its own result changes nothing. A **counted preview** (folders to create or reuse, cases to file or already filed) is shown before anything is written, and you can cancel. No new Azure DevOps call is made: the mapping the seed needs was already fetched. Seeding never overwrites your work. It only ever *adds* folders and files *unfiled* cases — your manual organization is safe. ## Bulk operations Tick cases by row to build a selection that **survives scrolling** and resets when you change folder or filter. **"Select all"** is bounded to the **visible, filtered results** — never the entire universe — so you always know exactly what you are acting on. Assign or remove a whole batch in a **single atomic write**, with an exact report afterwards (assigned / moved / unchanged). ## Readability: counts, flat view, name warnings * **Descendant-inclusive counts** — each folder shows how many cases it holds *including its sub-folders*, the way Xray does, so a parent folder is never misleadingly shown as empty. * **Flat view** — switch from the tree to a flat list of **every filed case with its full folder path**. Search applies, and your choice of view is remembered. * **Sibling name warning** — a **non-blocking** notice appears when two sibling folders share the same name (trimmed, case-insensitive), including duplicates a seed might have produced. Creating or renaming stays allowed — it is a heads-up, not a wall. ## "Not in any plan" (formerly "ADO orphans") The node that lists cases planned nowhere is labelled **"Not in any plan"** (French: *Non planifiés*). The earlier "ADO orphans" wording was too easy to confuse with "unfiled" — two different ideas. *Unfiled* means *you* haven't put it in a Repository folder yet; *Not in any plan* means **Azure DevOps** has it in no test plan. The definition didn't change, only the wording. ## Manual ordering Within a folder, reorder cases by **drag & drop** or with the **keyboard** (move up / move down). Each folder has a **sort mode**: `id` (default), `title`, or `manual`. The first time you reorder, the folder switches to manual sort — and switching back to an automatic sort asks for **explicit confirmation**, so a manual order is **never overwritten silently**. ## Related The concept, Unfiled, search and "used in". Turn your organized folder into a real ADO plan.