SPECIFICATION & BOUNDARIES
Supported bank statement formats.
TaxStatement accepts UTF-8 encoded CSV files and narrow, structured 4-column text PDF statements up to 4 MB or 2,000 transactions.
1. CSV specifications
- Encoding: UTF-8 (ASCII compatible).
- Delimiters: Comma (,), Semicolon (;), or Tab (\t). If using European comma decimals, choose semicolon or tab.
- Row limit: Up to 2,000 rows per file for deterministic, in-memory processing.
- Column mapping: Requires Date, Description, and either a Signed Amount (+ for inflow, - for outflow) or separate Debit/Credit columns.
2. Number & decimal conventions
Dot decimal (UK / US)
Example: 1,234.56 or -45.00. Comma is treated as thousands separator.
Comma decimal (Europe)
Example: 1.234,56 or 45,00-. Period is thousands separator. Use semicolon delimiter.
No guessing rule
Ambiguous or corrupted amounts halt the import rather than quietly being parsed as zero.
3. What is NOT supported
- Scanned documents or image-based PDFs (No OCR is performed).
- Multi-currency statements in a single file.
- Direct automatic tax calculations or HMRC online submissions.
- Automatic deletion of transactions (duplicates are flagged for human review only).