Inspiration

In the fast-paced freelance and vendor economy, manual escrow and delayed invoice settlements are major bottlenecks. Traditional payment platforms lack autonomous intelligence to verify contract milestones or resolve disputes fairly without human intervention. We built TrustLedger AI to bridge this gap, creating a secure, self-regulating micro-escrow web platform that combines automated payment flows with advanced AI reasoning.

What it does

TrustLedger AI functions as an autonomous vendor escrow and dispute mediator. It allows clients and vendors to lock milestone funds securely using the PayPal Sandbox API, tracks delivery hashes via secure SHA-256 logs, and leverages Google Gemini AI to calculate a dynamic TrustScore (0–100%). If a dispute arises, Gemini acts as an autonomous mediator, analyzing evidence logs and rendering fair verdict recommendations to release or hold funds automatically.

How we built it

  1. Backend & Logic: Python 5.0, Django 5.0, SQLite, and custom unit-test suites (46+ passing tests).
  2. Frontend & UI: Tailwind CSS structured in a sleek, dark-themed Bento-grid dashboard layout with interactive modals.
  3. APIs & Integrations: PayPal REST API (sandbox transactions) and Google Gemini API (Sentiment analysis, TrustScore generation, and dispute mediation).
  4. Security & Compliance: Strict .env credential isolation, global Django CSRF middleware protection, input sanitization, and a comprehensive enterprise legal framework (/legal/policies). ## Challenges we ran into
  5. API Fallback Handling: Ensuring the platform fails gracefully with clear user alerts when API keys are missing or network endpoints experience latency.
  6. Asynchronous Escrow Triggers: Designing clean data models to securely link dynamic PayPal transaction states with AI confidence thresholds ($>= 0.95$ auto-release rules). ## Accomplishments that we're proud of
  7. Achieving a fully responsive, enterprise-grade Bento-grid dashboard with zero critical console errors.
  8. Implementing an autonomous AI mediation engine that provides transparent reasoning logs for every transaction.
  9. Building a production-ready legal and security compliance package (including Terms of Service, Privacy Policy, Data Protection Policy, and AI Data Usage statements). ## What we learned We gained deep insights into structuring secure fintech web architectures, handling safe asynchronous payment webhook routing, and effectively utilizing Large Language Models for automated, legally compliant contract arbitration. ## What's next for TrustLedger AI
  10. Transitioning from PayPal Sandbox to live production merchant routing.
  11. Integrating multi-currency support and decentralized blockchain-backed immutable audit logs.

Built With

Share this project:

Updates

Submission history