Safe Work Pay

Demo

About

The problem it solves Freelancing platforms often suffer from **trust and payment issues**. Clients sometimes **refuse to pay** even after work is completed, and freelancers may **abandon tasks** after receiving payment upfront. This leads to **frequent disputes**, delays, and an overall **lack of fairness**. **SafeWorkPay** solves this by: * Using **Avalanche Fuji Testnet smart contracts** to hold payments securely in **escrow** until the task is completed. * Allowing clients to release funds **only when satisfied** with the submitted work. * Providing a **dispute resolution system** powered by **Eliza OS AI agent**, which analyzes the project requirements and submitted code. * In case of a dispute, the AI agent calculates the **completion percentage**, and the funds are **automatically split** between client and freelancer accordingly. This makes the entire freelancing process more **secure, transparent, and fair**, without the need for third-party moderators or lengthy manual review processes. ---

How it's made

Challenges we ran into One of the **biggest challenges** was integrating the **Eliza OS AI agent** with the blockchain-based system. Since Eliza OS runs as an external AI agent and the blockchain operates in a **trustless, isolated environment**, connecting the two in a **reliable and secure way** required careful architectural planning. Additionally, managing **dependencies and environment compatibility** between the **smart contract layer (Solidity + Avalanche Fuji)**, the **AI backend (Node.js + Eliza OS)**, and the **React frontend** led to some **complex bugs**—especially during testing and automation. We overcame these hurdles by: * Creating a **bridge layer** (using Node.js) to safely pass information between the AI agent and the smart contract. * Breaking the integration down into smaller, testable parts to isolate and resolve issues. * Thoroughly testing the AI logic with mock data before connecting it to real smart contract functions. These challenges were tough but helped us learn a lot about **cross-system architecture**, **AI integration**, and **real-world decentralized app development**. ---