CVM Planned Open Data API vs apicvm Today
Developers searching for a CVM open data API roadmap often mix three things: what dados.cvm.gov.br offers today, what CVM announced in its Plano de Dados Abertos (PDA) 2026–2028 (~July 21, 2026), and product APIs like apicvm that already ship a filings workflow.
This page separates planned official infrastructure from available product APIs — without claiming CVM has already released a ticker-first filings download API equivalent to apicvm.
What CVM announced (PDA 2026–2028)
In July 2026, CVM published its Plano de Dados Abertos 2026–2028, a strategic plan for expanding open data access. Among the initiatives, CVM describes work toward a public API on the open data portal.
Important framing:
- This is a roadmap, not a shipped product with a release date.
- Expected scope (based on public announcements) aligns with open data portal datasets and metadata — not necessarily a full document workflow with per-filing PDF download and text extraction.
- A future official API would be positive for the ecosystem — more programmatic access to regulator-published data benefits everyone.
What CVM open data provides today
Right now, dados.cvm.gov.br is primarily:
| Today | Description |
|---|---|
| Bulk datasets | CSV/ZIP downloads (financial statements, registries, funds, etc.) |
| CKAN portal | Package metadata, dataset catalog, periodic refreshes |
| CKAN API | Metadata and dataset listing — not a ticker-first filings product API |
You can list datasets and download bulk files programmatically via CKAN endpoints. That is different from:
resolve PETR4 → list DFP 2024 → download PDF → extract page markdownFor a deeper comparison of bulk open data vs product APIs, see CVM open data vs apicvm.
What apicvm provides today
apicvm is a shipping REST API (/v1) for B3 public company regulatory filings:
export APICVM_KEY='apicvm_...'
export APICVM_URL='https://apicvm.dev'
# Resolve ticker → list filings → download or extract
curl -H "Authorization: Bearer $APICVM_KEY" \
"$APICVM_URL/v1/companies/resolve?query=PETR4&by=ticker"
curl -H "Authorization: Bearer $APICVM_KEY" \
"$APICVM_URL/v1/documents?ticker=PETR4&type=DFP&year=2024&perPage=20"| Capability | CVM open data today | CVM planned API (roadmap) | apicvm today |
|---|---|---|---|
| Availability | Bulk portal + CKAN | Announced, not shipping | Shipping |
| Primary model | Dataset / ZIP download | Likely dataset-oriented (TBD) | Per-filing documents |
| Ticker-first list | No (join CSVs yourself) | Unknown | Yes |
| Original PDF download | Via bulk files, not REST doc ID | Unknown | GET /v1/documents/:id/file |
| Async markdown extraction | No | Not announced | Yes (callback) |
| Developer docs | Portuguese official portal | TBD | English /v1 contract |
| Auth | Public datasets (no key) | TBD | API key |
apicvm is not an official CVM product — it is a third-party API built over ingested CVM corpus data. CVM remains the regulator and authoritative publisher.
Critical distinction
CKAN metadata API ≠ filings product API.
Listing dataset packages on the open data portal does not replace:
- Resolving a B3 ticker to a company record
- Filtering documents by type (DFP / ITR / FRE) and year
- Downloading a specific filing PDF by stable ID
- Extracting page-level text for LLM pipelines
Do not assume the planned CVM API will replicate that full stack on day one. Strategic plans describe direction; product details, endpoints, and timelines are not fixed.
When to follow the CVM roadmap
| Scenario | Reasonable approach |
|---|---|
| You only need aggregate historical CSV panels | CVM open data bulk downloads today |
| You can wait for official infrastructure | Track PDA announcements; no guaranteed date |
| Portuguese-first, regulator-hosted datasets are enough | Open data portal when API ships |
| Academic cross-sectional research at scale | Bulk ZIP/CSV (see CSV vs API) |
When to use apicvm today
| Scenario | Reasonable approach |
|---|---|
| Agent or SaaS: "fetch PETR4 latest DFP" | apicvm now |
| Need stable document UUIDs and PDF download | apicvm now |
| RAG over regulatory PDF pages | apicvm async extraction |
English developer docs and versioned /v1 contract |
apicvm now |
| Production integration cannot wait on roadmap | apicvm now |
They may coexist
A future CVM public API and apicvm are not necessarily mutually exclusive:
CVM official API (future) → regulator-hosted datasets, ecosystem standard
apicvm (today) → ticker-first filings workflow, PDF + extractionMany teams will use official open data for research panels and a product API for application hot paths — same pattern as SEC bulk data vs EDGAR-oriented product APIs in the US.
Current limitations (apicvm)
- Not a mirror of the entire CVM open data catalog.
- Covers ingested DFP / ITR / FRE — not every dataset type on dados.cvm.gov.br.
- Paid API ($9/month Pro) with rate limits — open data remains free.
- Corpus sync follows ingestion — not real-time CVM publication.
Next steps
Ready to integrate?
Get an API key and start querying Brazilian CVM filings programmatically.