Trust Centre

How Lendable handles data and decisions

This page reflects the current implementation in this repository and summarizes how lending insights are produced.

Open Banking connectivity

Lendable connects to bank providers through TrueLayer after explicit user consent. We request access to accounts, balances, and transactions for affordability analysis.

Six-month transaction analysis window

The lendability flow is designed around reviewing up to six months of transaction history to assess income regularity, commitments, and risk signals.

Explainable scoring

The scoring engine calculates a lendability score and structured metrics, including affordability and red-flag checks, so results can be explained clearly.

AI summaries from sanitized inputs

If enabled, OpenAI receives a sanitized assessment payload (score bands, summarized flags, and selected metrics) to generate plain-language explanations for applicants.

Data minimization approach

Transaction data is used to produce the live report and user-facing insights. In this implementation, long-term storage is limited and primarily demo-oriented.

Current product maturity

This codebase includes prototype and demo patterns (including in-memory server stores and browser localStorage) and should be hardened before production use.

Important notice

This Trust Centre content is implementation-aligned documentation, not a legal certification statement. For production deployment, complete legal review, security review, and regulatory validation are required.