Brazil Government Procurement Scraper (PNCP / ComprasNet Tenders & Contracts)

Extract Brazilian public procurement data — tenders (editais/licitações), contract awards, and price-registration records — from the official PNCP (Portal Nacional de Contratações Públicas, the national portal mandated by Lei 14.133/2021, successor to ComprasNet). Filter by state (UF), modality, agency CNPJ, date range, and keywords. Output is clean, normalized JSON. For data engineers, market-intelligence teams, suppliers tracking opportunities, and researchers.

Built on PNCP's official public API (no scraping of login-walled pages), so results are stable and complete. Data is public under Brasil's Lei de Acesso à Informação and licensed CC BY 4.0 (commercial reuse allowed with attribution). No personal data is collected — see Compliance.

What it does

• Tenders (tenders): published procurement notices by publication date — object, agency, unit/UF, estimated value, modality, dates, status, legal basis.
• Contracts (contracts): signed contract awards — who won (legal-entity supplier), values, validity period, category.
• Open tenders (open_tenders): tenders currently accepting proposals.
• Normalizes Brazilian fields into a clean, English-keyed schema (PT values preserved): decoded sphere (Federal/Estadual/Municipal), UF + município + IBGE code, BRL values, ISO dates, and a direct PNCP link per record.
• Client-side keyword filter (accent/case-insensitive) over the object text — PNCP has no full-text search.

Input

Field	Type	Required	Description
dataType	enum	–	tenders (default) · contracts · open_tenders
dateFrom	string YYYY-MM-DD	for tenders/contracts	Start of the publication-date range
dateTo	string YYYY-MM-DD	✅	End of range (proposal horizon for open_tenders)
modalities	string[]	–	Lei 14.133 modality codes for tenders. 6=Pregão Eletrônico (default), 4=Concorrência Eletrônica, 8=Dispensa, 9=Inexigibilidade, …
uf	string	–	2-letter state filter, e.g. SP, SC
cnpjOrgao	string	–	Keep only one agency, by 14-digit CNPJ
keywords	string[]	–	Keep records whose object contains ANY term
maxResults	integer	–	Cap total records (default 500)
includeRaw	boolean	–	Attach the full original PNCP record under _raw (PII-stripped). Default off

Output (one record per Dataset item)

{
  "recordType": "tender",
  "pncpId": "83021808000182-1-000200/2026",
  "object": "CONTRATAÇÃO DE EMPRESA ESPECIALIZADA … TELECOMUNICAÇÃO …",
  "agency": { "cnpj": "83021808000182", "name": "MUNICIPIO DE CHAPECO", "sphere": "Municipal", "power": "Não aplicável" },
  "unit": { "name": "Prefeitura Municipal de Chapecó", "uf": "SC", "municipality": "Chapecó", "ibgeCode": "4204202" },
  "modality": { "id": 6, "name": "Pregão - Eletrônico" },
  "status": { "id": 1, "name": "Divulgada no PNCP" },
  "values": { "estimated": 1342860.0, "awarded": null, "currency": "BRL" },
  "dates": { "published": "2026-05-01T00:00:12", "proposalClose": "2026-05-19T08:55:01" },
  "sourceUrl": "https://pncp.gov.br/app/editais/83021808000182/2026/200",
  "_source": "Portal Nacional de Contratações Públicas (PNCP) — CC BY 4.0"
}

Contract records add supplier (legal-entity only — see Compliance), contractType, validity dates, and valorGlobal/valorInicial.

Example input

{ "dataType": "tenders", "dateFrom": "2026-05-01", "dateTo": "2026-05-31", "modalities": ["6"], "uf": "SC", "keywords": ["telecom", "internet"], "maxResults": 200 }

Data source, freshness & compliance

• Source: PNCP official consulta API. Licence: CC BY 4.0 — commercial reuse permitted with attribution (each record carries _source).
• Freshness: live per run (PNCP updates continuously).
• 🔒 No personal data (LGPD): only public procurement catalog data. Government contracts can be awarded to a natural person (tipoPessoa = PF); in that case the supplier's CPF and name are personal data and are stripped — the record marks supplier.redactedForPrivacy: true. Only legal-entity (PJ) supplier CNPJ + company name are returned. The publishing-system user field is never emitted.
• Tip: keyword queries scan more pages on the 50-records/page tenders endpoint — narrow with uf + a tight date range for speed. Very large pulls can take a few minutes; the actor throttles politely to respect PNCP's servers.

FAQ

What's the difference between PNCP and ComprasNet? ComprasNet (Compras.gov.br) is the legacy federal procurement portal. Under Lei 14.133/2021, publication migrated to PNCP — the national portal that aggregates federal, state and municipal procurement. This actor reads PNCP, so you get the broadest coverage (some legacy-only auctions may appear solely on ComprasNet).

Can I get licitações as Excel or CSV? Yes — results land in an Apify Dataset, which exports to CSV, Excel (XLSX), JSON, HTML or XML with one click (or via API). No parsing needed.

Is this data legal to use commercially? Yes. PNCP data is public under Brasil's Lei de Acesso à Informação and licensed CC BY 4.0 (commercial reuse with attribution — every record carries a _source field). The actor collects no personal data (LGPD): natural-person suppliers are automatically redacted.

How fresh is the data? Live — each run queries the official PNCP API directly. New tenders typically appear on PNCP within minutes of publication.

Can I monitor new tenders automatically? Yes — add an Apify Schedule (e.g., daily, with yesterday→today as the date range and your keywords/UF) and connect a webhook or integration (Sheets, Slack, Make, n8n) to the run results.

Maintenance

Weekly automated health-check; endpoint-change monitoring. Issues & requests via the Apify Store contact.

---

Published by ScrapeBR Studio — Brazilian public-data actors. Data © the Brazilian Government via PNCP (CC BY 4.0); this actor is an independent tool and is not affiliated with or endorsed by the Brazilian Government.