Access DFP and ITR Notes and Management Reports by Section Name

DFP (annual) and ITR (quarterly) filings from Brazilian public companies rarely arrive as one file. A single DFP cycle can include the standardized financial statements, explanatory notes, management report, fiscal council opinion, audit committee report, and director declarations — each a separate entry in the CVM document catalog.

If your pipeline needs footnotes or MD&A text, listing type=DFP&year=2024 without a name filter returns everything. apicvm lets you narrow to the annex you need before download or markdown extraction.

The problem

Financial analysis workflows often target specific annexes:

Annex	Why it matters
Explanatory notes	Segment revenue, contingencies, debt covenants, related parties
Management report (relatório da administração)	Narrative on margins, strategy, one-off items
Audit committee report	Internal control and audit oversight commentary
Fiscal council report	Brazilian governance body review of financials
Quarterly performance comments (ITR)	Interim MD&A updates between annual cycles

Without section filtering, you download large PDF packages and search manually — or run extraction on files you never query.

How DFP/ITR section names appear

CVM catalog name values for financial filings mix Portuguese labels and regulatory codes. Common patterns:

Demonstrações Financeiras — core financial statements
Notas Explicativas — explanatory notes to the financials
Relatório da Administração — management report
Parecer do Conselho Fiscal — fiscal council opinion
Relatório do Comitê de Auditoria — audit committee report
Declaração — director or officer declarations

Exact strings vary by issuer and filing year. Banks like ITUB4 file many PDF sections per DFP cycle — counts can be much higher than a typical industrial issuer. Use partial name matches and inspect list results.

How apicvm helps

The name parameter on GET /v1/documents performs a partial match on catalog document names.

Find DFP notes for PETR4

export APICVM_KEY='apicvm_...'
export APICVM_URL='https://apicvm.dev'

curl -H "Authorization: Bearer $APICVM_KEY" \
  "$APICVM_URL/v1/documents?ticker=PETR4&type=DFP&year=2024&name=Notas&perPage=20"

Review returned name values — pick the notes annex explicitly:

{
  "data": [
    {
      "id": "660e8400-e29b-41d4-a716-446655440001",
      "type": "DFP",
      "year": 2024,
      "name": "Notas Explicativas às Demonstrações Financeiras",
      "dateRef": "2024-12-31"
    }
  ]
}

Find management report

curl -H "Authorization: Bearer $APICVM_KEY" \
  "$APICVM_URL/v1/documents?ticker=RADL3&type=DFP&year=2024&name=Relatório da Administração"

ITR quarterly comments

curl -H "Authorization: Bearer $APICVM_KEY" \
  "$APICVM_URL/v1/documents?ticker=VALE3&type=ITR&year=2024&name=Comentário&perPage=20"

ITR filings use dateRef for quarter-end dates — combine year and dateRef when multiple quarters return.

Python: notes and MD&A fetcher

import os, requests

BASE = os.environ["APICVM_URL"]
H = {"Authorization": f"Bearer {os.environ['APICVM_KEY']}"}

FINANCIAL_SECTIONS = {
    "notes": "Notas",
    "management_report": "Relatório da Administração",
    "audit_committee": "Comitê de Auditoria",
    "fiscal_council": "Conselho Fiscal",
}

def find_financial_sections(ticker: str, doc_type: str, year: int) -> dict[str, list]:
    sections = {}
    for key, name_prefix in FINANCIAL_SECTIONS.items():
        resp = requests.get(
            f"{BASE}/v1/documents",
            params={
                "ticker": ticker,
                "type": doc_type,
                "year": year,
                "name": name_prefix,
                "perPage": 20,
            },
            headers=H,
        )
        resp.raise_for_status()
        sections[key] = resp.json()["data"]
    return sections

sections = find_financial_sections("RADL3", "DFP", 2024)
for theme, docs in sections.items():
    for doc in docs:
        print(f"{theme}: {doc['name']} ({doc['id']})")

Queue POST /v1/document-text-extractions only for the sections your footnote parser or RAG index needs.

Footnote analysis workflow

Earnings quality and segment analysis depend on notes, not just the face financial statements:

1. GET /v1/documents?ticker=RADL3&type=DFP&year=2024&name=Notas
2. Pick notes document ID
3. POST /v1/document-text-extractions → receive page markdown via callback
4. Query: segment revenue mix, tax contingencies, non-recurring items, lease adjustments
5. Cross-check management report for narrative on same themes

See extract financial footnotes from DFP for prompt patterns once text is indexed.

Banks and high file counts

Large banks file many DFP sections per cycle. Example pattern from ITUB4 integrations:

# List all DFP 2024 files first to discover names
curl -H "Authorization: Bearer $APICVM_KEY" \
  "$APICVM_URL/v1/documents?ticker=ITUB4&type=DFP&year=2024&perPage=50"

# Then filter to financial statements or notes
curl -H "Authorization: Bearer $APICVM_KEY" \
  "$APICVM_URL/v1/documents?ticker=ITUB4&type=DFP&year=2024&name=Demonstrações"

Use perPage=50 (maximum) and paginate with page= when exploring unfamiliar issuers.

Current limitations

No structured footnote parser — apicvm returns page-level markdown, not tagged note objects like Note 12 — Segment Information. Your pipeline segments and classifies text within pages.
Name strings are issuer-specific — partial matches help, but first-time issuers may require a discovery list call without name.
Notes may span multiple files — some companies split notes across annexes; review all matches.
ITR quarter disambiguation — use dateRef when listing ITR to pick the correct quarter.
Corpus coverage — documents must be ingested in Hold; verify with a list call.

Next steps

Ready to integrate?

Get an API key and start querying Brazilian CVM filings programmatically.

Read the API docs Get API key