Inspiration

As freelancers, we often face a frustrating reality: when we send a $2,000 invoice to an international client, we mentally calculate our earnings as $2,000. But when the money actually hits our bank account, it is significantly less—often 5% to 8% less. Between cross-border transaction fees and hidden FX markups, we silently lose a huge chunk of our hard-earned money. To make matters worse, freelancers never know the exact net amount until the transaction is complete, and they have no idea if they should withdraw the funds immediately or wait for a better exchange rate. We built RateMate to solve this exact pain point.

What it does

RateMate is an AI-powered financial agent that acts as a "pre-invoice guard" for freelancers.

  1. Net Payout Prediction: Before sending an invoice, a freelancer inputs the amount. RateMate uses live FX rates and fee structures to predict the exact net amount they will receive in their local currency.
  2. Reverse Calculator: Freelancers can input their desired net amount (e.g., ₹50,000) and RateMate tells them exactly how much to invoice their client.
  3. Agentic AI Global Transfer Optimizer: Compares multiple platforms (Wise, Remitly, OFX) to find the best net payout and fastest delivery.
  4. ML FX Forecasting: Uses Machine Learning to predict the next 3-day exchange rate trend and recommends the best time to withdraw.
  5. Smart FX Alerts: AI suggests the best target rate and notifies the user when the market hits it.
  6. AI Copilot: A 24/7 floating chatbot that answers financial questions in natural language.

How we built it

We built the backend using Python and FastAPI, integrating the PayPal Developer APIs and Frankfurter FX APIs:

  • PayPal APIs for authentication and transaction fee structures.
  • Frankfurter API for real-time and historical FX rates.
  • Supabase (PostgreSQL) for secure user authentication and database management.
  • Groq (Llama 3) for generating natural language recommendations and drafting reminders.
  • Scikit-Learn for training the Machine Learning model for FX forecasting.

The frontend is built with HTML5, CSS3, and Vanilla JavaScript, using AG Grid to create a powerful, interactive ledger table. The entire application is deployed on Render.

Challenges we ran into

  1. Cross-Currency Calculations: Handling complex cross-rates (e.g., EUR to INR) accurately while accounting for different provider fees required deep mathematical logic and real-time API integration.
  2. AI Prompt Engineering: Making sure the AI's financial advice was accurate, concise, and didn't hallucinate numbers. We had to fine-tune the prompts to ensure it clearly explained why it recommended a certain payout timing.
  3. State Management: Ensuring the dashboard updated dynamically without refreshing the page, especially when handling multiple currencies and live FX rates.

Accomplishments that we're proud of

We are incredibly proud of creating a fully functional, end-to-end prototype within the hackathon timeframe. Successfully integrating Agentic AI, Machine Learning forecasting, and real-time database management into a single, seamless dashboard was a massive achievement. The AG Grid integration and the premium dark-theme UI make the platform feel like a production-ready SaaS product.

What we learned

We gained a deep understanding of how cross-border payments work, especially the hidden costs associated with FX conversion. We also learned how to effectively use AI not just as a chatbot, but as an analytical decision-making engine that processes financial data and provides actionable advice.

What's next for RateMate

In the future, we plan to:

  • Integrate multi-currency wallets to allow freelancers to hold funds until the FX rate is highly favorable.
  • Add support for more alternative payment gateways (like Wise or Payoneer) to provide a comprehensive comparison.
  • Build a mobile app for instant push notifications when the optimal FX rate hits.
Share this project:

Updates

Submission history