apicvm.dev
Download llms.txt Get API key

LlamaIndex Tools for Brazil CVM Filings

Building LlamaIndex CVM filings workflows? Treat apicvm as an HTTP tool layer: resolve → list → download/extract, then index markdown or PDFs in LlamaIndex. This is not a managed LlamaIndex connector.

The problem

Agents that cite Brazilian filings need:

  • Stable document UUIDs for citations
  • Filters by ticker, type (DFP/ITR/FRE), and year
  • Page-level text without scraping the CVM portal

LlamaIndex handles retrieval; apicvm handles regulatory file access.

Tool sketch (Python)

import os, requests
from llama_index.core.tools import FunctionTool

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

def list_filings(ticker: str, doc_type: str = "DFP", year: int = 2024) -> str:
    r = requests.get(
        f"{BASE}/v1/documents",
        headers=H,
        params={"ticker": ticker, "type": doc_type, "year": year, "perPage": 20},
        timeout=60,
    )
    r.raise_for_status()
    rows = r.json()["data"]
    return "\n".join(f"{d['id']}\t{d['name']}" for d in rows)

list_tool = FunctionTool.from_defaults(fn=list_filings)

RAG path

  1. List documents for the coverage universe
  2. Download PDFs with GET /v1/documents/:id/file or request async page markdown via POST /v1/document-text-extractions
  3. Index chunks in LlamaIndex with metadata {ticker, type, year, document_id}
  4. Cite document_id in answers

For a fuller pipeline see Build a RAG pipeline and LangChain tools.

Async extraction note

Extraction is not synchronous. You register a callback URL, receive 202, then wait for the webhook. Do not tell the agent "wait on a status endpoint" — that endpoint does not exist in v1.

Current limitations

  • No official LlamaIndex integration package from apicvm
  • Student keys cannot extract text (403)
  • Corpus is not "all Brazilian issuers"

Next steps

Ready to integrate?

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

Read the API docs Get API key