elias569/fintech-app

This repository is a prototype platform that connects startups looking for funding with investors, using AI to help match the two sides. Startups upload their pitch deck as a PDF, and investors describe what kind of companies they like to fund, called their thesis, along with sectors and stages they focus on. The platform then works out matches in both directions, and an AI service drafts a personalized introduction email meant to reduce the awkwardness of a first outreach message. The project is split into a FastAPI backend and a Vite based frontend, connected through a documented set of API endpoints, including one to upload a pitch deck, one to check an startup's investor matches, one for investors to register their preferences, one to see an investor's current dealflow, and one to generate an intro email between a specific startup and investor. The backend currently stores data in memory rather than a permanent database, which the README frames as fine for prototyping but something that would need to change for real use. Interactive API documentation is available through Swagger once the backend is running. Login on the frontend is described as a local demo only, with no real backend authentication: any non-empty name, email, and password will work, and the role picked at sign in, either startup or investor, decides which dashboard a user lands on. The session is kept in the browser's local storage until logged out. Running the backend needs Python with FastAPI installed and an OpenAI API key placed in an environment file, since the matching and email generation rely on GPT. Running the frontend just needs Node and npm to install packages and start the development server. An optional Redis cache can be added to speed up the AI based match scoring, caching results for five minutes, but the app works without it too. The README does not mention a license.