- Schema: account_id FK on checks and layout_fields; UNIQUE per-account on check_no and field_name
- DB: runtime migration recreates both tables to add account_id (assigns existing rows to account 1)
- Routes: GET /api/accounts lists all; GET /api/account/:id replaces hardcoded id=1; POST /api/account/setup always creates a new account and returns accountId
- checks.js: all queries scoped by account_id; POST requires account_id in body
- pdf.js: resolves account from check's account_id instead of id=1; layout fields fetched per-account
- import-mdb.js: always INSERTs a new account (never deletes existing); all records tagged with new accountId
- Frontend: account switcher in header; activeAccountId persisted to localStorage; all API calls pass account_id; switching accounts reloads checks; wizard and import auto-switch to newly created account
- All checks shown in one table by default (load all, no server-side printed filter)
- Add payee search, date-from/to, and status filter controls to toolbar
- Filtering is client-side; no extra API calls on filter changes
- All checks get Edit/Delete buttons regardless of printed status
- All checks get checkboxes for PDF selection
- Remove separate Reprint button and reprintCheck function
- Remove printed guard from PUT and DELETE endpoints
steve