ynab-import-transactions

Files (1)

• SKILL.md

SKILL.md (first lines)

# Bulk-import transactions from a CSV/export

**Tools:** `list_accounts`, `get_transactions`, `modify_transactions`

Only run this on an explicit request — it creates transactions in bulk.

## Steps

1. Resolve `plan_id` and target `account_id` (see [setup](../README.md#setup);
   call `list_accounts` if the account isn't specified).
2. Map each source row to: `date` (ISO), `amount` (milliunits, negative for
   outflows), `payee_name`, `memo`. Leave `category_id` unset — categorize
   afterward with `ynab-triage-inbox`.
3. Generate a stable `import_id` per row (max 36 chars, e.g.
   `YNAB:<amount>:<date>:<index>`) so re-running this on overlapping data doesn't
   duplicate.
4. Cross-check with `get_transactions` (scoped to the account/date range) for
   same date+amount+payee overlaps the `import_id` might miss.
5. Preview the mapped rows and total count; flag likely duplicates.
6. On confirmation, call `modify_transactions` `action="create"` in batches of
   ~100.
7. Report created vs. skipped counts.

## Done when

Every intended transaction exists exactly once, with no duplicates or drops.

## Note

Future-dated rows are rejected by YNAB — use `modify_scheduled_transactions`
instead.

From lukendatigh/ynab-mcp-server, MIT.

More from lukendatigh/ynab-mcp-server