Inspiration

What if your financial history existed, but the system couldn't see it?

Millions of people are financially active yet invisible to traditional credit systems. Immigrants arriving in the United States, freelancers earning irregular incomes, and individuals in developing economies can demonstrate financial responsibility without having the conventional credit history lenders expect.

Coming from Belarus and studying in the United States made me question why someone's financial credibility should depend so heavily on where they live and whether they have previously borrowed money.

We built FOLIO — Credit Beyond Your Credit History to explore a different approach: evaluate the financial evidence people already have, rather than penalizing them for the credit history they don't.

What it does

FOLIO is an alternative financial assessment platform that turns everyday financial records into an explainable, portable financial passport.

The experience follows four steps:

  1. Connect: Import financial records through CSV, PDF statements, or supported banking integrations.
  2. Understand: AI helps interpret transactions, identify income patterns, and flag information requiring review.
  3. Assess: A transparent, rules-based assessment evaluates available financial evidence, showing contributing factors, confidence, and limitations.
  4. Export: Generate a professional financial passport in PDF format that summarizes financial standing and its supporting evidence.

Unlike a black-box credit score, FOLIO lets users inspect and correct their information. When evidence is insufficient, it explicitly says so instead of inventing a misleading score.

Our demonstrations include financial profiles from both the United States and Belarus, showing how the approach can accommodate different financial environments.

How we built it

We developed FOLIO using Next.js, React, TypeScript, Node.js, and Express, with a responsive interface inspired by the clarity and simplicity of modern financial applications.

Three integrations power its core functionality:

Google Gemini: Uses structured AI outputs to extract information from financial documents, suggest transaction classifications, and provide evidence-grounded explanations. Numerical assessments remain deterministic rather than relying on AI-generated calculations.

Capital One Nessie API: Connects FOLIO to a simulated banking environment, allowing users to import and inspect transaction evidence from sandbox accounts.

Tiger Data: Powers persistent financial records and historical cash-flow analytics using PostgreSQL, TimescaleDB hypertables, and continuous aggregates.

Additional technologies include PDFKit for passport generation, Zod for data validation, and Playwright for automated testing.

We used OpenAI Codex for implementation and coordination, with Claude Code independently reviewing code and identifying issues.

Challenges we ran into

Building trust into financial AI. An AI-generated explanation can sound convincing even when the underlying evidence is incomplete. We separated AI-assisted interpretation from deterministic calculations, added source references, and required human review of uncertain classifications.

Working with incomplete financial records. Banking transactions alone cannot establish someone's complete financial condition. We implemented coverage checks and explicit unscored outcomes rather than treating missing data as zero.

Integrating multiple services. Connecting Gemini, Nessie, and Tiger Data required handling different data structures, provider limitations, persistence, and error conditions without silently fabricating results.

Designing for accessibility and clarity. Financial assessments can be intimidating. We prioritized understandable explanations, responsive layouts, transparent evidence trails, and a polished PDF experience.

Accomplishments we're proud of

We created more than a static credit-score demonstration: FOLIO offers an end-to-end workflow from financial evidence ingestion and correction to an explainable assessment and exportable passport.

We successfully exercised real Gemini and Nessie integrations, connected persistent Tiger Cloud analytics, and tested how assessments behave when information is incomplete.

Most importantly, we designed a system that doesn't confuse missing credit history with financial irresponsibility.

What we learned

Building FOLIO reinforced that trustworthy financial technology requires more than sophisticated AI.

Explainability, data provenance, user control, and knowing when not to produce a score are just as important as generating insights.

We also learned how to combine generative AI with deterministic financial logic, integrate banking sandboxes, and build persistent analytics across multiple services.

What's next

Our goal is to develop FOLIO into an interoperable financial identity platform for people underserved by conventional credit infrastructure.

Next steps include expanding supported financial data sources, strengthening privacy and authentication, conducting fairness and validation studies, and exploring partnerships with financial institutions willing to evaluate alternative evidence.

A traditional credit score tells you how someone has borrowed. FOLIO explores what their broader financial behavior can tell you.

FOLIO is an experimental financial assessment tool, not a validated credit-risk prediction model or a lender-approved credit score.

Built With

Share this project:

Updates

Submission history