CredMate

CredMate is an AI-powered voice-based loan matchmaker that connects borrowers with lenders instantly through smart, human-like conversations.

Comment

💳 CredMate – AI-Powered Voice-Driven Loan Matchmaker

Inspiration

We were inspired by the struggles faced by first-time borrowers, immigrants, and underserved communities in accessing fair and understandable loan options. The lending landscape is often opaque and predatory. We asked: "What if getting a loan felt like speaking with a smart, friendly advisor who already understood your needs?" This led to CredMate — a conversational, AI-powered lending platform designed to make borrowing more human, transparent, and accessible.

What it does

CredMate is a voice-enabled loan matchmaking platform that simulates the end-to-end lending journey. Borrowers interact with a voice or chat agent to provide key information. An autonomous matcher agent negotiates with simulated lenders, presents tailored loan offers, and allows users to negotiate further. Approved users receive a full breakdown in a mock wallet dashboard — all in real time.

How we built it

  • Frontend: Built using React.js and Tailwind CSS for a mobile-first, responsive chat and wallet interface.
  • Voice Interface: Integrated with Vapi Voice Assistant Agent to support natural, real-time borrower interaction.
  • Agent Logic: Used Fetch.ai’s Transact Framework to model contract-based communication between borrowers and decentralized lender agents.
  • Backend Simulation: Loan agents were emulated with randomized constraints. Supabase was used to store session data and mock analytics.

Challenges we ran into

  • Adapting Fetch.ai’s Transact model for a fully frontend-simulated environment.
  • Designing a clear and robust voice UX that gracefully handles vague or unclear input.
  • Balancing realistic loan logic while ensuring local-only, safe financial simulations.
  • Creating fallback flows for edge cases, such as when no matching loan is available.

Accomplishments that we're proud of

  • Successfully built a complete lending simulator with agentic communication and autonomous matchmaking.
  • Delivered a fully functional voice assistant and wallet interface.
  • Created a human-like borrowing experience that emphasizes trust and simplicity.
  • Seamlessly integrated voice and chat UX into a responsive, polished frontend.

What we learned

  • How to design and simulate decentralized agent negotiation in a financial context.
  • The impact of intuitive voice UX in reducing user friction and boosting trust.
  • The importance of clarity, trust, and transparency in building AI-powered financial tools.
  • How to architect scalable, agent-based systems for real-world use cases.

What's next for CredMate

  • Transition from simulated agents to live Fetch.ai economic agents on the mainnet.
  • Integrate live credit scoring APIs like Experian or Plaid.
  • Expand to support multi-party loan comparisons (e.g., top 3 lenders).
  • Launch multilingual support and voice biometric security features.
  • Implement KYC/AML protocols for real-world readiness.
  • Open-source our agent matcher and wallet UI as a FinTech developer toolkit.

Built With

Share this project:

Updates