Belvo
Belvo is an open finance data and payments platform for Latin America, serving banks, lenders and fintechs in Brazil and Mexico. Apps built on it have connected more than 80 million accounts, it has processed and analyzed more than 3 billion transactions, and it handles about $1 billion a year in payments. Nubank, Inter, Banamex, Banco Plata, Smart Fit, Adyen, Alipay, iFood and inDrive appear among its users. Belvo does not lend or take deposits itself.
On the data side, Belvo pulls consented banking, employment and fiscal data and turns it into credit inputs: machine learning categorization of raw transactions, recurring expense detection, account and income verification, Risk Insights features for lenders' models, an employment score for job stability and an income estimator. Open Finance Score, developed by FICO, runs neural networks over thousands of transactions across a person's accounts and returns a score with reason codes, reporting an AUC of 74% and a KS of 36%, to complement bureau scores for thin file applicants.
On the payments side, Belvo offers direct debit and managed collections in Mexico and one tap biometric Pix payments, smart collections and transfers in Brazil, through a payment institution registered with Banco Central do Brasil. Inter built Belvo's data infrastructure into its credit journey, lifting connection success from 62% to 76%. A free sandbox is open to developers, and production plans start at $1,000 a month.
Capability Axes
Capability grades
15 of 15 axes rated · 11 graded A or B
Models carry two core workflows. Machine learning cleans and classifies raw transactions into the categories, recurring expenses, income and risk features lenders use, and FICO's neural network Open Finance Score turns consented transactions into a credit score with reason codes. Account connectivity, employment data retrieval and payment initiation work without them, so the AI layer sits on the credit side of a platform that also moves money.
Scores, reason codes and categorized features go back to the lender, which decides. Institutions can feed the data into an existing score or internal model, and consumers control access through consent. Accept, decline and escalation thresholds sit with the lender, and no review route of Belvo's own is described for its model outputs.
Open Finance Score reports an AUC of 74% and a Kolmogorov Smirnov statistic of 36% and returns reason codes with each score. Risk Insights exposes the categorized inputs so a lender can inspect them and build its own models, as Monet did in a technical study of features drawn from transaction data. FICO's validation of the score is not mapped to a supervisory model risk standard.
Inter built Belvo's data infrastructure into its Brazilian credit journey and lifted successful connections from 62% to 76%, and Aro halved first installment defaults in nine months after building Belvo's banking data into its own rules and eight proprietary models. Both results are named and measured, but they come from data connectivity and the customers' own models rather than Belvo's categorization or the FICO score. Belvo also claims twice the loan approvals for credit use cases, and Monet's and Aplazo's chief operating officers are quoted.
The categorization models are trained by Belvo on a large volume of historical transactions. Enrich Your Data does not store what it receives, and the connection service can delete user derived data when a configured period runs out. Whether customer data can train models shared across products, and how long the FICO model or any inference supplier keeps data, are not stated.
The end user privacy policy covers financial data handling and names the controlling client and the local Belvo entities. Customers configure how long links and their data are kept, and Enrich Your Data does not store submitted data. Protections vary by product and connection settings, and a full supplier and retention schedule for every model workflow is not available.
Belvo carries ISO 27001:2022 certification for its information security management system and complies with the card industry's data security standard, hosts on Amazon Web Services and describes encryption, software testing and managed platform access. Supporting material sits in its trust center, and no current certificate opens from the public site.
Belvo's Brazilian payment institution is listed by Banco Central do Brasil as a payment institution operating as a payment transaction initiator, which covers the local payments business. Belvo does not lend or take deposits, and lenders such as Inter run their credit decisions under their own licenses, using Belvo's data and scores as inputs.
Open Finance Score is pitched at people with thin or no credit history, which makes fair lending central to its use. FICO, its developer, is described as a pioneer in responsible and explainable AI, and the score returns reason codes, but neither Belvo nor FICO offers a fair lending procedure, disparate impact analysis or adverse action workflow for it.
Reason codes accompany each score so a lender can review the analysis, and requests about a person's underlying data go to the institution responsible for it. The lender controls approval and its own review process. Belvo offers no compensation for a wrong score.
FICO develops the Open Finance Score model, and Belvo builds its own categorization and enrichment models. Amazon Web Services hosts the platform. The inference chain behind the rest of the platform, including any outside model providers, is not named.
Documented APIs, a Connect Widget and a Connect Page reach banks, employment records and fiscal data across Brazil and Mexico, and Belvo MCPs connect AI agents to the same open finance data. Inter wired Belvo's infrastructure directly into its credit journey, combining open finance and employment data with its own process. A free sandbox covers connections and consent, and payment initiation runs on Pix in Brazil and direct debit in Mexico.
The enterprise platform runs on Amazon Web Services, and customers set how long links and their data are stored. A committed hosting region for enterprise deployments, or a region choice for the scoring service, is not stated.
A free sandbox with test data lets developers evaluate the API, and production access starts at $1,000 a month on the entry plan. Larger deployments are scoped with sales, and the price list carries no separate charge per score or integration.
Lenders, banks and fintechs in Brazil and Mexico use Belvo for credit, income verification, account services and collections. Risk Insights, the employment score and Open Finance Score serve credit teams, including thin file lending, while direct debit, Pix and managed collections serve businesses collecting recurring payments, such as Smart Fit. Each market has its own connections and consent flows.
Alternatives to Belvo
Other products that do the same kind of work as Belvo, and what each one does.
Tink, a Visa company, sells open banking software to European banks, lenders, and fintechs. Its lending products turn bank transaction histories into inputs for affordability and credit reviews.
Yodlee, owned by STG since September 2025, sells open finance data and analytics to lenders, banks, fintechs, and wealth platforms.
CreditXpert helps mortgage lenders and brokers see how an applicant's mortgage credit score could change before the loan is priced or underwritten.
Klavi turns permissioned Brazilian Open Finance data into income, spending and credit risk signals. A borrower consents to sharing information from bank accounts through a connection managed by Klavi Instituição de Pagamento.
Ntropy turns raw bank transactions and statements into structured data for lenders, fintechs and finance teams.
Moneyhub is an open finance platform from Bristol that helps banks, lenders, pension providers and wealth managers turn consented financial data into customer insight, lending decisions and payments.
Pricing
Vendor-published figures are labeled as such. Figures labeled “Estimated” are derived from third-party sources and have not been confirmed by the vendor.
| Entry Price | Pricing Basis | Data Protection Terms | Implementation | Source |
|---|---|---|---|---|
|
$1,000 a month for the entry production plan
$1,000 baseline
|
A free sandbox covers evaluation with test data, and production starts at $1,000 a month on the entry plan. Larger deployments are scoped with Belvo's sales team. | End user financial data is handled under Belvo's service privacy policy, and customers set retention for account links and their data. | Implementation covers the consent flow and connections to bank, employment, fiscal or payment data. The sandbox is free, and a standard implementation fee is not listed. | Vendor Published |
Belvo's entry production plan costs $1,000 a month for access to its open finance data and payments platform, after a free sandbox. The plan price covers platform access, with no separate charge listed per score.