Can AI replace Tiller?

Build a closest honest personal substitute for Tiller in an empty repository.
Use Next.js 15, TypeScript, SQLite, Drizzle ORM, and local-first encryption; do not offer alternative stacks.
The core loop is: import financial CSV transactions into a transparent spreadsheet model, categorize them with rules and templates, reconcile balances, and export clean reports without moving money.
Make the first run work locally with one documented command.
Store all user data locally by default and make export straightforward.
Put secrets in .env, ship .env.example, and never commit credentials.
Support CSV and OFX import with saved column mappings and duplicate detection.
Implement accounts, transactions, transfers, categories, payees, notes, and split transactions.
Add deterministic categorization rules whose matches can be inspected and undone.
Provide reconciliation, monthly budget, cash-flow, net-worth, and export views.
Keep an immutable import log and require explicit confirmation before destructive changes.
Encrypt local data, create automatic backups, and document restore on a fresh machine.
Include clear empty, loading, success, and recoverable error states.
Add input validation, safe filenames, and graceful handling of unavailable APIs.
Write focused tests for the core transformation and one end-to-end happy path.
Create a README with setup, architecture, permissions, data location, and backup steps.
Do not add accounts, billing, telemetry, analytics, or a hosted control plane.
Do not claim to reproduce proprietary data, network liquidity, regulated access, or frontier infrastructure.
Deliberately leave out initiating payments or moving money.
Deliberately leave out tax, payroll, or accounting compliance guarantees.
Deliberately leave out screen-scraped bank access and commercial institution aggregation.
Finish by running the tests and listing the exact commands used.