Pull Asset Code, Preço, and IRRF from an XP or Rico Nota de Corretagem Screenshot
Most trading tools treat a Portuguese-language brokerage note like any other document — then miss the IRRF day-trade column and mistake PETR4 for a plain ticker without its class suffix. Semantic extraction reads B3-regulated fields by meaning, not label position.
5-10s per page · Up to 99% accuracy on printed text
What You Can Pull from an XP or Rico Nota de Corretagem
XP and Rico follow the B3-mandated SINACOR format — every brokerage note shares the same regulated layout. With Custom Column Extraction, you name the fields and the AI locates each by its financial meaning.
PETR4, VALE3, WINFUT — Brazilian asset codes encode both the company and share class in the numeric suffix. General OCR sees a string; the AI sees a structured identifier.
Shares or contracts traded — appears beside the asset code in the main operations table. The AI distinguishes it from price by number format and column context.
Per-unit fill price in BRL. The nota labels this "Preço de Ajuste" — the AI reads it by column heading, not by pixel position, adapting to minor variations in the rendered PDF.
Buy or sell indicator — a single letter that determines tax treatment. Venda (sell) rows trigger IRRF and capital gains tracking; Compra (buy) rows establish cost basis.
Imposto de Renda Retido na Fonte — withheld at 1% of net gains for day-trade or 0.005% of sale value for swing-trade. The AI distinguishes the two by reading the row's classification marker.
Exchange, clearing, and registry fees charged by B3 — packed into a dense table at the bottom of the nota. The AI extracts each line item into its own column, separating blended values into clean cells.
Why a Standardized Brokerage Note Still Trips Up Generic Tools
XP and Rico's nota de corretagem uses the B3-mandated SINACOR format — same layout every time. But the fields inside use Portuguese financial terminology that general-purpose OCR was never built to understand.
Portuguese Labels Confuse General OCR
"Preço de Ajuste", "Emolumentos", "IRRF D", "C/V" — these are standardised labels in the SINACOR format, but a generic tool returns them as raw Portuguese text. It doesn't know that "IRRF D" means day-trade tax withholding while "IRRF" without the D means swing-trade. The label language alone creates a semantic barrier.
Asset Codes Carry Hidden Meaning
PETR4 is not just "Petrobras" — the "4" means ordinary shares (ordinárias), while PETR3 would mean preferred shares (preferenciais). Brazilian asset codes encode both company and share class in a single string. A trading journal that logs only the ticker loses the class distinction needed for corporate action tracking and dividend calculations.
Dense Fee Table Merges Multiple Charges
Every nota de corretagem includes a fee summary at the bottom — emolumentos B3, taxa de liquidação, ISS, PIS, COFINS — packed into a compact table with small print and minimal row separation. Traditional table extraction collapses these into one undifferentiated block, losing the per-item breakdown needed for tax reconciliation.
Semantic Reading by Financial Role
Define IRRF Day-Trade and IRRF Swing-Trade as separate columns. The AI reads the nota's row classification — the "D" marker — and routes each IRRF value to the correct column. It knows that "IRRF D" is 1% of net daily gains and "IRRF" is 0.005% of sale value, because it understands what each label means in the B3 context.
Brazilian Asset Code Awareness
The AI recognizes PETR4 as "Petrobras, ordinary share class 4" — it understands that the numeric suffix is a share class indicator, not part of a quantity or price. Define one Ativo column and the full identifier, company name, and class are available in your spreadsheet without manual splitting.
Structed Fee Extraction by Line Item
Define columns for each B3 charge — Emolumentos, ISS, PIS, COFINS. The AI reads the fee table row by row and unpacks each value into its named column. What generic OCR returns as one messy block becomes five clean numbers, ready for your tax reconciliation workflow.
The COIR open-source project, built specifically to handle XP and Rico notas, highlights the same pain points its author encountered in 2017 — manual extraction that "até então todo esse trabalho era realizado de forma manual" (until then all this work was done manually) — confirming that automated extraction from brokerage notes remains a real, unsolved problem.
From a Nota de Corretagem to a Tax-Ready Spreadsheet
If you're consolidating trades from XP or Rico into your trading journal or monthly DARF calculation, here's what the extraction workflow looks like:
Upload the Nota de Corretagem
Drop the PDF or screenshot of your XP or Rico brokerage note into the upload area. The nota can be downloaded directly from the broker's portal as a SINACOR PDF, or captured as a screenshot from your phone. PNG, JPG, or PDF — the tool accepts all formats. Upload one full nota or batch an entire month's worth.
Name Your Trading Journal Columns
Define the fields you need: Ativo, Quantidade, Preço de Ajuste, C/V, IRRF Day-Trade, IRRF Swing-Trade, Emolumentos. Add a Computed Column like Net Proceeds (Valor - IRRF - Emolumentos) to calculate the net amount during extraction — no post-processing in Excel needed.
Export a Unified Spreadsheet
Processing takes about 5-10 seconds per page. The output is one XLSX or CSV file with a row per trade and the columns you defined as headers. Export directly to Excel for your existing trading journal template, or send to Google Sheets via the add-on. One brokerage note, one batch, one spreadsheet — no manual typing.
When Nota de Corretagem Extraction Works Best — and What to Watch For
When It Works Best
Standard SINACOR-format PDFs or screenshots. Machine-rendered text from XP or Rico's official nota reaches up to 99% field accuracy — the regulated layout helps rather than hinders.
Full-document notas showing header, operations table, and fee table. The entire nota in one frame gives the AI all the context it needs to correctly classify each field.
Cross-broker batch processing. The same column schema works across XP, Rico, and other SINACOR-standard brokers, plus US platforms like Robinhood and Coinbase in the same export.
When to Be Cautious
Screenshots that crop the bottom fee section. Emolumentos and IRRF totals live in the fee summary at the bottom of the nota. A cropped screenshot misses these field — capture the full page.
Compressed images forwarded through messaging apps. Small print in the fee table and asset codes degrade first when compressed — spot-check these fields if the image was forwarded via WhatsApp.
Image-based PDFs from non-standard formats. If the nota was scanned as an image rather than rendered as text (common with older brokers), the AI still extracts the fields — but accuracy depends on scan quality and the fee table may require verification.
Frequently Asked Questions
Can I extract Ativo, Preço de Ajuste, and IRRF directly from an XP or Rico nota de corretagem screenshot or PDF?
Yes. Define columns for Ativo, Quantidade, Preço de Ajuste, C/V, IRRF, and Emolumentos. The AI reads each field by semantic understanding of the SINACOR-standard format — it doesn't need an exact pixel match because it understands what each label means in the B3 context. The same column schema works whether the source is a native PDF downloaded from the broker portal or a screenshot taken on your phone.
How does the AI distinguish IRRF for day-trade from IRRF for swing-trade on the same nota?
XP and Rico follow the SINACOR standard, which marks day-trade operations with a "D" indicator in the operations table. The IRRF on these rows is calculated at 1% of net daily gains. Swing-trade rows show IRRF at 0.005% of the total sale value. Define separate columns like IRRF Day-Trade and IRRF Swing-Trade — the AI reads the row classification on the original document and routes each value to the correct column automatically. If a single nota mixes both trade types, each row is processed according to its own classification without manual separation.
What does the C/V indicator mean in the output — and does it affect the extraction?
C/V stands for Compra (buy) or Venda (sell) — it's a single-letter column that tells you the direction of each trade. The AI extracts it as a label alongside the numeric fields. This indicator is important for tax purposes: only Venda (sell) rows trigger IRRF withholding and capital gains tracking. Compra rows establish cost basis but don't generate tax events. In your trading journal or DARF calculation, the C/V column lets you filter by direction without cross-referencing the original document.
Can I batch XP and Rico brokerage notes alongside US broker trade confirmations in one export?
Yes. Define one set of column names — Ativo, Price, Quantity, Trade Date, Platform — and the AI processes XP nota de corretagem PDFs, Robinhood trade confirmations, and Coinbase transaction screenshots using the same definitions. Each document becomes one row in the output spreadsheet regardless of which country or regulatory system generated it. This is the core advantage of semantic extraction: you define the fields you need once, and the AI adapts to each platform's unique format automatically, whether it's a B3-regulated nota or a US brokerage confirmation.
Can I extract Emolumentos and individual B3 fees from the dense fee table at the bottom of the nota?
Yes. The fee summary at the bottom of every XP/Rico nota includes Emolumentos (B3 exchange fees), Taxa de Liquidação (clearing and settlement), ISS (municipal service tax), PIS, and COFINS. These are packed into a compact table with small print. Define separate columns for each fee line — the AI reads the table row by row and unpacks each value into its named column. For the most reliable result, upload the full nota without cropping the bottom section. If you need to reconcile fees across multiple traders, batch all monthly notas together and the AI will extract the same fee structure from each one.
Deep dives into trading and investment screenshot extraction: General screenshot extraction guide with a dedicated trading and investment logs section · How to define custom columns for any app UI including brokerage screens
For extracting data from other trading platforms, see the Trading & Investment Screenshots hub — covers Robinhood, Coinbase, Binance, Webull, eToro, and more.