💡 Inspiration
Every GitHub commit represents time, effort, and valuable development work. However, freelance developers often have to manually review their contributions, describe completed work professionally, calculate billable amounts, and create invoices for clients.
I wanted to solve this problem by connecting the tools developers already use with the tools they need to get paid.
That's how CommitPay was born — an AI-powered GitHub-to-PayPal invoicing assistant designed to simplify the journey from completed code to professional invoices.
🚀 What is CommitPay?
CommitPay transforms GitHub development activity into professional, AI-assisted invoice descriptions and helps developers create PayPal invoices.
Its goal is simple:
Your Code. Your Work. Your Payments.
⚙️ How I Built It
I built CommitPay using a React frontend, FastAPI backend, GitHub REST API, Ollama, and PayPal Developer Sandbox.
The application follows a straightforward workflow:
1. Connect GitHub Activity
Users provide a public GitHub repository and retrieve its commit history through the GitHub REST API. They can review and manually select the commits representing their completed work.
2. AI-Powered Work Descriptions
CommitPay uses Ollama with Meta's Llama 3.2 (3B) model to transform selected commit messages into clearer, professional descriptions of development deliverables.
The AI assists with writing descriptions, while the developer remains responsible for reviewing and approving them.
3. Build an Invoice
Developers manually enter their billable hours and hourly rate. CommitPay calculates the invoice total deterministically, ensuring that AI does not decide working hours, rates, or payment amounts.
4. Save Invoice Drafts
Users can save their invoice drafts locally and continue working on them later.
5. PayPal Integration
Using PayPal's Invoicing REST API and OAuth 2.0 authentication, CommitPay can create invoice drafts in the PayPal Developer Sandbox after explicit user confirmation.
No real payments are processed, and invoices are not automatically sent to clients.
🧠 Where AI Makes a Difference
Instead of using AI to make financial decisions, CommitPay uses it where it provides genuine value: converting technical development activity into understandable descriptions of completed work.
The model runs locally through Ollama, making the AI component local-first rather than dependent on a paid external AI API.
🛠️ Challenges I Faced
One of the biggest challenges was integrating three different systems: GitHub, a locally running AI model, and PayPal's invoicing platform.
I had to handle GitHub API limitations, implement PayPal authentication, coordinate frontend and backend communication, and ensure that invoice creation remained under the user's control.
Another important challenge was designing the AI workflow so that generated descriptions could assist developers without allowing AI to determine financial information.
📚 What I Learned
Building CommitPay helped me gain practical experience with:
- Integrating third-party REST APIs.
- Implementing OAuth 2.0 authentication.
- Building full-stack applications with React and FastAPI.
- Running and integrating local AI models using Ollama.
- Working with PayPal's Developer Sandbox and Invoicing API.
- Designing safer AI-assisted financial workflows.
- Testing and debugging a multi-service application.
This project also helped me understand how AI can improve everyday developer workflows when combined with existing platforms.
🎯 What's Next?
I envision CommitPay becoming a useful invoicing companion for independent developers and freelancers — reducing administrative work so they can spend more time building and less time preparing invoices.
CommitPay — Your Code. Your Work. Your Payments.
Built With
- artificalintelligence
- fastapi
- githubapi
- invoicing
- javascript
- llama3.2
- oauth2.0
- ollama
- paypal
- python
- react
- restapi
- three.js
- typescript
- vite