Sabot
Your third-party safety layer for transactions
Created on 8th January 2026
•
Sabot
Your third-party safety layer for transactions
The problem Sabot solves
Sabot addresses the fundamental lack of verifiable proof in today's online economy. Currently, a vast amount of economic activity—over $700 billion in peer-to-peer trade and millions of freelance transactions—relies on fragile, easily forged evidence.
-
Sabot acts as a "safety layer" for peer-to-peer and freelance transactions, solving the problem of relying on easily forged screenshots and "good faith".
-
It uses AI to verify user identities and automatically converts informal chat logs into binding, tamper-proof smart contracts.
-
The platform integrates a blockchain escrow system that locks funds (e.g., ETH) and only releases them once both parties confirm the delivery of goods or services.
-
Sabot acts as a digital auditor by analyzing transaction context for fairness, such as warning users if a listing price is suspiciously below the market average.
-
Resolves disputes fairly in cases where a deal goes wrong by using neutral arbiters and oracles can review the immutable blockchain record to resolve the dispute and issue refunds or releases.
-
Provides live collaborative drafting where users can create contracts and have AI monitor the context for fairness and consistency. Once both parties confirm and "sign" the deal, the document is locked to prevent modification and permanently uploaded to the blockchain as an immutable record.
Challenges we ran into
One major technical hurdle we faced was a race condition in our escrow smart contract, where near-simultaneous confirmations from different users caused inconsistent states or reverted transactions. We resolved this by re-architecting a stricter state machine within the contract, implementing ReentrancyGuard to prevent parallel execution exploits, and adding frontend safeguards that disable conflicting actions while transactions are pending.
Additionally, building the real-time collaborative features presented significant integration challenges. We initially struggled with compatibility issues between yjs, WebSockets, and Next.js, which were compounded by limited documentation. To overcome this, we pivoted from a traditional WebSocket server to a serverless approach using PartyKit. This architectural shift resolved the connection stability issues and allowed us to deploy a seamless, low-latency collaborative drafting environment without the overhead of managing complex socket infrastructure.
Tracks Applied (1)
Winner Classification
Technologies used