← Back to home PT
Honest comparison

B3 data APIs — an honest comparison

B3 fundamental-data APIs — an honest comparison

Verified on 2026-09-01, on the competitors' own pages. This comparison is maintained by Dados B3 itself — read it with that discount and check the links.

Three rows moved against us in this revision (the previous one was dated 2026-08-10): bolsai now ships a native MCP connector, publishes its methodology, and prices essentially at our level. The old table understated it on three counts, and this is the second most fetched page by AI assistants. Re-dating without re-checking would have kept the error in our favour.

What changed on 2026-09-03, and again against us. The bolsai price we published (R$ 49/mo) was out of date: their Pro tier is R$ 129/quarter, or R$ 43/mo, against our R$ 49.90. They are cheaper, and the previous table said we were level. Two houses that were missing also came in — Partnr and HG Brasil — and the second forces a category correction: their own FAQ says the API does not provide fundamental indicators or financial statements, so it does not compete here, it competes on quotes.

That is the second revision in a row where the error was in our favour.

Anyone who needs Brazilian stock (B3) company data via API has a few real options today. Each has genuine strengths; the choice depends on what you value.

Dados B3 brapi bolsai Partnr HG Brasil Fintz Economatica
Entry price R$ 0 (200/day) · Pro R$ 49.90/mo R$ 0 (15,000 req/mo) · Startup R$ 119.99/mo R$ 0 (200/day) · Pro R$ 43/mo (R$ 129/quarter) 7-day trial · price not public free tier · price not public R$ 12,500/mo (only public tier) on request (institutional)
Fundamentals 17 years (2010–) since 2009 (Pro) 11–15 years since 2010, 160+ indicators not provided (their own FAQ) since 2010 since 1986
Public methodology Yes — every formula, account by account No Yes — published, cites CVM account Claims "auditable formulas", does not publish them No No Inside the system, for subscribers
Published quality tests Yes — invariant suite; if it fails, nothing ships No Self-benchmark on blog No No No No
Point-in-time multiples (no look-ahead) Yes — price of the 1st session on or after the actual filing No No Not mentioned No Own base for backtest (enterprise) Not announced
Restatement history Yes — 9,467 recorded No No Not mentioned No No No
Open page per company Yes — 356, no key No No No No No Inside the system
No-signup trial Full WEGE3 4 tickers Playground 7-day trial free tier Limited test key No
AI connector (MCP) Yes Yes Yes (uvx bolsai-mcp) No No No Yes

Where each is better than us — honestly

  • brapi: near-real-time quotes, crypto, currencies and community volume. If you need intraday prices, use brapi — we only have end-of-day (COTAHIST), by legal and scope choice.
  • bolsai: cheaper than us — R$ 43/mo on the quarterly plan against our R$ 49.90 — the same free-tier shape (200/day, no card), a native MCP connector installable in one line, a published methodology, and prices back to 1986 against our 2010. On price history they win by 24 years. If a long quote series is what you need, go there. They also publish 300+ users and testimonials; we publish none.
  • Partnr: 160+ indicators, and they go beyond what we cover — value added statements, registration forms, insider transactions and corporate events, with Python and JavaScript SDKs. If breadth of fields is what you need, it is there. The price is not public, which is a cost for anyone who wants to decide on their own.
  • HG Brasil: a mature ecosystem since 2009 — quotes, currencies, bitcoin, REITs and dividends, with a free tier. It does not compete on fundamentals: their FAQ answers that the API provides no fundamental indicators and no financial statements. It is in the table so that nobody comes looking for that here believing otherwise.
  • Economatica: unmatched institutional depth (since 1986, ready-made analysis screens, dedicated support). If your fund pays R$ 2k+/mo, it's a different product category.
  • Fintz: enterprise infrastructure with SLA and a base built for institutional backtesting.

Where none of the five does what we do

This list SHRANK in the 2026-09-01 revision, and the item that left was the first one: a public methodology stopped being ours alone — bolsai now publishes its own, citing the regulator's account codes. Ours remains a differentiator of DEGREE (a formula per indicator, published before the number), not of kind.

  1. Invariant tests — dozens of automatic checks (the balance sheet ties, the income statement ties, the tax identity, price coverage) run before each publication. If one fails, the update doesn't ship.
  2. Point-in-time multiples — P/E computed with the price you COULD have paid (1st session from the day the report was published), not the Dec-31 price matched with a report that only came out in March. For backtesting, that's the difference between a real result and an illusion.
  3. Restatements never silently corrected — when a company refiles a report, both versions stay on record.

Sources

Prices and features verified on 2026-08-10: brapi pricing · bolsai · Fintz · Economatica · ours: pricing, methodology, trial. Found something outdated? Reach us via the portal.