Load Brazil CVM Filings in R
Analysts who live in R still need primary filings — not only scraped tables. This guide shows an R CVM filings API workflow with httr2: resolve a ticker, list documents, download PDFs.
The problem
R packages that parse CVM ZIP/CSV dumps help for accounting panels. They do not give you:
- Ticker-first document lists with stable IDs
- Section-level FRE PDFs filtered by name
- A single auth pattern for production jobs
Install and auth
install.packages("httr2")
library(httr2)
base <- Sys.getenv("APICVM_URL", "https://apicvm.dev")
key <- Sys.getenv("APICVM_KEY") # from /signup
apicvm_get <- function(path, ...) {
req <- request(paste0(base, path)) |>
req_headers(Authorization = paste("Bearer", key)) |>
req_url_query(...)
resp <- req_perform(req)
resp_body_json(resp)
}Resolve and list DFP 2024 for VALE3
company <- apicvm_get("/v1/companies/resolve", query = "VALE3", by = "ticker")
str(company)
docs <- apicvm_get(
"/v1/documents",
ticker = "VALE3",
type = "DFP",
year = 2024,
perPage = 20
)
ids <- vapply(docs$data, function(d) d$id, character(1))
names <- vapply(docs$data, function(d) d$name, character(1))
data.frame(id = ids, name = names)Filter FRE risk sections
risks <- apicvm_get(
"/v1/documents",
ticker = "VALE3",
type = "FRE",
year = 2024,
name = "FatoresRisco",
perPage = 20
)Download one PDF
download_filing <- function(doc_id, dest) {
request(paste0(base, "/v1/documents/", doc_id, "/file")) |>
req_headers(Authorization = paste("Bearer", key)) |>
req_perform(path = dest)
}
if (length(docs$data) > 0) {
download_filing(docs$data[[1]]$id, "vale3-dfp.pdf")
}vs pandas
If your team splits R and Python, the same endpoints work in both. See Load Brazil CVM filings into pandas.
Build a simple table of filings
docs_to_df <- function(payload) {
data.frame(
id = vapply(payload$data, \(d) d$id, character(1)),
name = vapply(payload$data, \(d) d$name, character(1)),
type = vapply(payload$data, \(d) d$type, character(1)),
stringsAsFactors = FALSE
)
}
dfp <- docs_to_df(apicvm_get(
"/v1/documents",
ticker = "PETR4", type = "DFP", year = 2024, perPage = 20
))
print(dfp)Auth headers
Authorization: Bearer is preferred. X-API-Key also works. Keep keys in .Renviron, not in committed notebooks. See CVM API authentication.
When R packages that parse ZIPs still help
Local parsers remain useful for bulk accounting panels from Dados Abertos CSVs. Use apicvm when you need a specific issuer PDF or FRE section by ticker. Comparison context: CVM CSV dumps vs API.
Current limitations
- apicvm returns metadata and files — not a tidy accounting DataFrame of every DFP line.
- Extraction to page markdown is async (Pro credits), not an R sync helper.
- Empty lists usually mean “not in corpus yet,” not “issuer has no filings.”
Next steps
Ready to integrate?
Get an API key and start querying Brazilian CVM filings programmatically.