Inspiration

Informal digital agreements made over chat apps frequently suffer from vague definitions, forgotten terms, and a complete lack of structural trust. We built TradeProof AI to turn loose conversations into explicit, immutable agreements without forcing users to draft heavy legal contracts from scratch.

What it does

TradeProof AI parses conversation text to automatically isolate core deliverables, pricing models, and key dates. It proactively flags missing or ambiguous agreement criteria for manual resolution. Once both the seller and the customer confirm the finalized terms, the system canonicalizes the data, creates a unique salt, and anchors a 32-byte SHA-256 hash onto the network. Users can download a private JSON copy of the transaction receipt and drop it into an offline verification tab to immediately validate that not a single character has been modified since it was signed.

How we built it

The system is engineered as an offline-first desktop prototype using a Node.js Express backend, a PostgreSQL relational database managed via Prisma Client, and a fast React frontend bundled with Vite. Cryptographic handling is conducted completely client-side in the browser using the native Web Crypto API.

Challenges we faced

Our biggest hurdles were dealing with external rate limits and connectivity dropouts during live third-party API testing. We resolved this gracefully by embedding a robust, deterministic local mock architecture into our test server boundaries, allowing full workflow simulation and strict data verification without network dependencies.

What we learned

We learned the power of explicit RFC 8785 JSON canonicalization to ensure identical cryptographic hashes across completely separate environments, showing how local processing can preserve user data privacy perfectly.

Built With

Share this project:

Updates

Submission history