Dados do Brasil em uma única API.

Since mid-2026 Brazil’s tax authority issues CNPJs with letters (e.g. 12ABC34501DE35). Old regex (^\d{14}$) rejects valid customers — signups break silently.

Validates both formats with the official check-digit algorithm, normalizes, formats, and generates test batches so QA covers the new format before it becomes a production incident.

Misclassifying a transaction’s IBS/CBS code means a rejected invoice or, worse, a tax assessment. The official tables change without warning on your radar.

Full official classification table with every code linked straight to the exact statute article, plus the complete NCM table (10,515 codes) — both re-imported every month.

Every rejected e-invoice means manual rework, shipping delays, and at volume, a whole team just chasing down XML errors field by field.

Real validation against the complete official government XML schema (not a shallow check) — returns the exact field, the violated rule and a fix suggestion, before you ever submit.

The official price list ships as a raw monthly spreadsheet with no stable naming — parsing, cleaning and versioning it eats a developer’s time every month.

Complete official list (26,001 items) with regulatory registration, barcode and regulated prices already adjustable by state tax rate — re-imported from the source every month.

Billing with the wrong procedure code is a classic reason claims get denied. Matching two code tables by hand, item by item, doesn’t scale.

Free-text semantic search → procedure code, plus an official regulator crosswalk where one exists (always labeled as such, never disguised as a guess) and a text-similarity suggestion only when no official mapping exists.

Verifying that a provider is real, active, and pulling its tax ID/address/coordinates for a network or map means cross-referencing public registries in incompatible formats.

Complete, active national base — all states, every facility type — with name, tax ID, address and coordinates, searchable by state, city or facility type.

The official INSS/IRRF brackets and the 2,400+ CBO occupations come as raw portarias and loose tables — re-keyed by hand in every payroll system, and wrong versions cause fines.

The current CBO list (canonical + synonyms) and the 2026 INSS/IRRF brackets with their legal basis, plus an auditable progressive INSS and IRRF calculator with the 2025-2026 tax relief.

Looking up the right CEST for a product, or the interstate ICMS rate between two states, means checking CONFAZ annexes and state legislation — by hand, every time.

The complete CEST×NCM table (1,013 rules, CONFAZ) plus the interstate ICMS rates between every pair of states, with the rule that produced each rate.

Every health app ends up embedding a partial, outdated CID-10 table — and the official DATASUS files are raw, unindexed and hard to query.

The complete Brazilian CID-10 (categories + subcategories, pt-BR), searchable by free text with accent-insensitive matching, plus chapter navigation and code lookup.

The Central Bank publishes the raw series, but applying the official correction methodology (which months count, pro-rata, mid-period index changes) is where everyone errs.

Official monthly series (SELIC, IPCA, INPC, IGP-M, IGP-DI, INCC) plus a correction calculator that applies the standard official methodology and shows every factor used.

Foreign developers quoting prices in Brazil need the official PTAX rate, but scraping the Central Bank or parsing historical series is fragile and undocumented.

Official PTAX rates for USD, EUR, CHF and ARS with 36 months of history, latest-quote endpoint and a BRL↔currency converter that uses the correct rate for any date.

Validating a CNPJ is not enough — you need the full firmographic record. The official Receita Federal dataset is huge, poorly documented and painful to maintain.

The complete public company record from the Receita Federal open data: legal name, trade name, CNAE (primary + secondary), status, address, capital, shareholders and Simples/MEI — refreshed automatically from the official source.

Knowing a company's shareholders isn't enough — you need the whole network. Who else does this person control? What is the full economic group?

The corporate graph built from 28M shareholder records × 69M companies: shareholders of a CNPJ, every company of a shareholder, and the complete economic group behind any company.

Finding qualified Brazilian leads means scraping registries and praying the contacts are right. Nobody sells you the full national base with phone and email ready to import.

72.8M establishments searchable by CNAE, UF, municipality, size and status — with phone and email — plus batch enrichment and one-click CSV export for your CRM.

Risk-scoring Brazilian companies usually needs paid bureau data. The basic signals — status, age, size, capital, Simples — are all public, but scattered.

Transparent risk score (status + age + size + capital + regime), sector benchmarks by CNAE/state (capital and age percentiles), and full headquarters/branches structure — computed from the official base.

Validating a healthcare provider means checking the CNES registry and the company registry separately, in incompatible formats.

The 494,662-facility national health registry crossed with the full CNPJ base: unit type, location, coordinates, and the company's legal data in one response.

KYC isn't a one-time check. Regulators require continuous monitoring, but detecting changes in the company registry requires versioned data most APIs don't keep.

A watchlist of CNPJs with snapshot-based change detection: status, capital and shareholder-count changes between monthly versions of the official base.

A company's registered capital in 2010 isn't comparable to today's — you need it restated by official inflation, and doing that per company by hand is tedious.

The registered capital of any CNPJ restated by IPCA or SELIC using the official Central Bank methodology, with the exact factor applied.

You have a name — not a CNPJ. Finding which companies a person controls means combing through registry files name by name. Nobody exposes this as an API.

Search 28M shareholder records by name and get every company the person is linked to — with status, size, capital and state. Plus the recovered CPF and a health-professional flag: if the person works in healthcare (CNES), you see their occupation, council and CNS. The reverse of Company Graph: start from the person.

Verifying a doctor, nurse or therapist — where they work, what they do, whether they are in the SUS — means digging through DATASUS files in a proprietary compressed format nobody documents.

The national health workforce: 4M+ unique workers and 6.7M work links across every healthcare facility, searchable by name — with occupation (CBO), professional council, recovered CPF and every establishment they work at — refreshed monthly from the official source.

A doctor who owns a clinic, a nurse with a side business, a therapist who is a partner in a chain — knowing the business interests behind a health professional means cross-referencing two government databases by hand.

Every company linked to a health professional, crossed by their CPF: their own MEI/EI (legal name ends with the CPF) and companies where they appear as a shareholder — with status, size, capital and state.

Sizing a healthcare market — "how many physiotherapists are there?" — means downloading and parsing the entire national registry, state by state, in a proprietary format.

Aggregated workforce analytics: total workers and unique professionals per state, SUS participation, and a searchable occupation directory (CBO) with real counts — the demand side of Brazilian healthcare.

CMED lists maximum allowed prices, not what anyone actually pays. Clinics and hospitals buying gloves, reagents or drugs have no public benchmark of real transaction prices.

791K+ real purchases from the Ministry of Health price bank (BPS, 2009-2023): medicines AND medical supplies, with the actual unit price paid, the supplier (CNPJ + name), the buying institution, date and state — searchable, comparable by supplier, with yearly history.

Brazilian health software must code diagnoses (CID), procedures (TUSS/CBHPM/SIGTAP), drugs (ATC) and exams (LOINC) — each from a different official table, each in a different format, each updated on its own schedule.

ATC/DDD (WHO, 7K drugs with defined daily dose), CIAP-2 (686 primary-care rubrics), CBHPM/TUSS (32.9K private procedures/OPME/medications) and LOINC-BR exam names — all in one searchable API with versioned sources.

SIGTAP is the official SUS table — procedures, medications and OPME with the exact amounts the government pays — but it ships as a monthly fixed-width file dump that no one wants to parse and keep current.

The complete unified SIGTAP table (DATASUS, current month): 5,023 procedures with SH/SA/SP values, complexity, sex/age rules, financing, plus 82K CID-10 links, 195K CBO occupation links, services, habilitations, bed types, modalities, RENASES, SIA/SIH mappings and compatibility rules — searchable in milliseconds.

Private health billing in Brazil runs on TUSS codes, but ANS publishes 65 tables in scattered formats — and each code has a validity window you must respect when submitting TISS claims.

Every TUSS table (procedures, daily rates/taxes/gases, medications, OPME, domains) with the official validity dates of each code (inicio/fim vigência) — one searchable API, updated 2x/month from the ANS official platform.

Brazil is migrating from CID-10 to the WHO ICD-11 — a completely different code system. Every health software will need the new codes and a way to translate legacy CID-10 data.

The official ICD-11 MMS linearization (WHO 2024-01): 35,310 categories, 788 blocks, 28 chapters — searchable with hierarchy navigation — plus the official WHO ICD-10↔ICD-11 mapping files (12.3K + 17.2K links) to translate codes both ways.

Every medicine in Brazil needs an ANVISA registration, but the official open-data file is a messy CSV in a server with broken TLS — nobody parses it and keeps it current.

The official registry, clean and searchable: 43,444 products with active ingredient, therapeutic class, registration number, status (Active/Inactive), holder (CNPJ + name) and expiry — plus 16,395 usage/prescription restrictions and a live cross-reference to CMED prices.