About dadosbrasil
dadosbrasil publishes the open data of Brazil's National Registry of Legal Entities (CNPJ) in three formats: HTML (this site), Markdown, and a JSON API. We don't collect anything ourselves: every row comes directly from the public files on the Brazilian federal open-data portal.
Sources
- Receita Federal — CNPJ: company registration, establishments, partners, Simples Nacional enrollment, reference tables (CNAE, legal nature, municipalities, countries, partner qualifications).
- IBGE: municipality names and codes.
Refresh cadence
The Federal Revenue publishes a new snapshot of the CNPJ archive roughly once a month. dadosbrasil is fully re-imported on every release — we don't do incremental diffs, since the official consistency guarantee is per snapshot. So the effective "as-of" date of our data is the date of the most recent RFB snapshot, not today; companies changed very recently may take up to ~6 weeks to show up.
Methodology
Pages are generated straight from the RFB files with no manual enrichment. The only additional computations are:
- Company network (
/cnpj/.../network): groups companies sharing at least one individual partner (by masked CPF), the same zip code, or the same legal representative. Only first-degree connections are counted. - Partner network (
/cpf/.../network): starts from a CPF, lists the companies where it appears as a partner plus each one's first-degree network. - One-line addresses: server-side concatenation of street-type + street + number + complement + neighborhood fields.
Privacy and CPF masking
The Federal Revenue itself publishes individual partners' CPF numbers in the masked form ***XXXXXX** — the first three and last two digits are always hidden. dadosbrasil just reproduces this form. We never try to reconstruct the full CPF, nor to cross-reference masked CPFs with other databases to de-anonymize anyone.
If you show up as a partner and want your page (/en/cpf/***XXXXXX**) removed, the first step is to speak with the Federal Revenue — the source. Since our data mirrors the public snapshot, any RFB correction propagates automatically on the next monthly cycle. For urgent cases, reach out via GitHub Issues.
Known limitations
- Companies marked as "Cancelled" stay in the database — they are historical RFB records and don't imply current activity.
- The "Share Capital" field is the declared value; it's not automatically updated over time.
- Legal-entity partner names are sometimes blank when the source doesn't fill the field.
- English translation applies only to the scaffolding (labels, descriptions). The data itself (legal names, addresses) stays in Portuguese, as published.
License and attribution
The original data is public and free to use under the Federal Revenue's terms. Derivative works should cite Receita Federal as the source and, optionally, dadosbrasil as the intermediary/reformatter.
Source code
dadosbrasil is open source at github.com/brazildata/api.