The import page includes a Bank Preset dropdown with pre-configured column mappings for popular banks (Chase, Wells Fargo, Bank of America, Capital One, and more).

How presets work:

1. Select your bank from the dropdown
2. The date, description, and amount column names are automatically filled in
3. Upload your file and proceed to preview

If your bank isn't in the list, SenticMoney will attempt to auto-detect columns when you upload a file. If auto-detection doesn't find the right columns, you can use AI Format Detection or enter column names manually.

When you upload a CSV or Excel file that doesn't match a known bank preset, the "Analyze with AI" button becomes available. This uses Google Gemini AI to identify which columns contain dates, descriptions, and amounts.

How it works:

1. Upload your bank's CSV or Excel file
2. If columns aren't auto-detected, click "Analyze with AI"
3. Review the AI's detected column mappings in the result dialog
4. Click "Apply & Save Profile" to use the mappings and save them for future imports

When AI detects a bank format and you click "Apply & Save Profile", the column mappings are saved locally on your computer. On future imports:

• Auto-match: When you upload a file, SenticMoney compares the column headers against saved profiles. If there's a match, the columns are filled in automatically and you'll see a notification.
• Manual selection: You can also pick from the Learned Profiles dropdown to apply a saved profile.

Profiles are identified by a fingerprint of the column headers, so they work even if the file name changes. Profiles are stored locally in app/static/data/learned_bank_profiles.json and are never sent to any external service.

PDF bank statements are processed entirely on your computer using pdfplumber. No data is sent externally.

How it works:

1. Upload a PDF bank statement
2. Tables are extracted from the PDF pages
3. Columns (date, description, amounts) are auto-detected from the extracted table headers
4. Preview the transactions and confirm the import

Limitations:

• Works best with statements that have clearly formatted tables
• Scanned/image-based PDFs are not supported (the PDF must contain selectable text)
• Some banks use non-standard layouts that may not extract correctly
• If your PDF doesn't import cleanly, try downloading a CSV or OFX version from your bank instead

Columns not detected correctly

Open your file in Excel or a text editor to see the exact column header names. Column names are case-sensitive. Enter them exactly as they appear in the file.

Dates parsing incorrectly

Different banks use different date formats (MM/DD/YYYY vs DD/MM/YYYY vs YYYY-MM-DD). If dates look wrong in the preview, try a different bank preset or use AI detection which identifies the date format automatically.

Split amounts (separate debit/credit columns)

Some banks use two columns for amounts (e.g., "Withdrawals" and "Deposits" or "Debit" and "Credit"). Click "Use Split Columns" on the import page and enter both column names.

CSV file has extra header rows

Some banks include account information at the top of CSV files before the actual column headers. SenticMoney auto-detects these metadata rows and skips them. If it doesn't work correctly, try removing the extra rows manually in Excel.

OFX/QFX file won't parse

Some OFX files have malformed XML tags. SenticMoney preprocesses common issues (like empty tags), but severely corrupted files may not import. Try downloading a fresh copy from your bank.