ProofPay AI

AI clarifies freelance tasks, verifies submitted work against agreed requirements, and connects client approval to PayPal Sandbox payments for fairer digital work.

Comment

Inspiration

Freelancing is becoming increasingly accessible to students and beginners, but small digital projects often start with unclear requirements. A client may ask for a poster, design, or other deliverable without clearly defining what is expected. This can lead to misunderstandings, repeated revisions, and payment disputes.

We wanted to build a platform that makes small freelance projects more transparent by bringing clear agreements, AI-assisted work review, and digital payments into one workflow.

What is ProofPay AI?

ProofPay AI is an AI-powered trust platform for small freelance tasks.

It helps clients and freelancers establish clear expectations before work begins and review submitted work against those expectations.

Our workflow includes:

  1. Create a task: The client describes the work, budget, deadline, and expected deliverables.
  2. Generate an agreement: AI helps convert the task description into structured deliverables and acceptance criteria.
  3. Approve requirements: The client reviews and approves the agreement.
  4. Submit work: The freelancer uploads completed work and supporting files.
  5. AI-assisted verification: The system compares the submission with the agreed requirements and highlights potentially missing or unclear items.
  6. Client review: The client makes the final decision to approve the work or request revisions.
  7. PayPal integration: PayPal Sandbox supports the payment checkout and capture workflow.

AI provides assistance, not an automatic final decision. The client remains responsible for reviewing and approving submitted work.

How we built it

We built ProofPay AI using a React and TypeScript frontend with a Python FastAPI backend.

  • Frontend: React, Vite, TypeScript, Tailwind CSS
  • Backend: Python, FastAPI
  • Database: PostgreSQL with SQLAlchemy
  • AI: Ollama with Qwen3 1.7B for agreement generation and advisory work verification
  • Payments: PayPal REST API Sandbox
  • Deployment: Vercel for the frontend and Render for the backend
  • Database hosting: Neon PostgreSQL

We also implemented file uploads and PDF text extraction to help evaluate submitted work.

The payment integration uses a $1.00 USD Sandbox test transaction. This is a technical demonstration, not real-money processing or an escrow service.

Challenges we faced

One major challenge was designing a workflow where AI could help evaluate work without taking control away from the client.

We also faced challenges integrating PayPal Sandbox order creation, approval, and capture while handling payment states safely and preventing duplicate captures.

Another challenge was deploying a local AI model in a cloud environment. Our initial implementation uses Ollama locally, so making AI functionality available in production requires a hosted inference connection.

Connecting the FastAPI backend to a persistent PostgreSQL database and deploying the frontend and backend separately were also important learning experiences.

What we learned

Building ProofPay AI helped us understand how AI can be integrated into a real application rather than used only for text generation.

We gained practical experience with REST APIs, database integration, file handling, AI-assisted evaluation, payment workflows, environment variables, GitHub, and cloud deployment.

Most importantly, we learned that trust in digital work requires more than AI. Clear expectations, transparent reviews, and explicit human approval are equally important.

What's next?

We plan to improve hosted AI inference, strengthen the verification process, improve the user experience, and expand payment and project-management capabilities.

Our goal is to make small freelance projects clearer, safer, and easier to manage for both clients and independent creators.

Built With

Share this project:

Updates

Submission history