What I Built
I built FinLens, a small AI-powered web app that helps people understand financial news without needing a finance background.
A lot of financial news is written for people who already understand terms like inflation, interest rates, tariffs, bond yields, GDP, or central-bank decisions. For everyone else, headlines can feel important but confusing.
I built FinLens for the friend who sends me a financial headline and asks:
“Okay, but what does this actually mean?”
You choose a date and a region — such as India, the US, China, Europe, Asia-Pacific, or Global, and FinLens finds a few stories that are worth understanding and explains them in simple language.
The goal isn't to create another news feed or summarize entire articles. It is to provide just enough context to make the news click.
Demo
Try FinLens here:
https://finlens-d8r2.onrender.com
Choose a date and region, then explore the stories FinLens thinks are worth understanding.
Code
The project is open source on GitHub:
https://github.com/urgetolearn/finlens
How I Built It
FinLens is built around Gemma, using Ollama's cloud model access with gemma4:cloud.
The application is built with Python and Streamlit. It retrieves financial news from RSS sources, retrieves relevant article content when needed, and passes the information through an AI agent that is instructed to focus on understanding rather than simply summarizing.
The agent's job is to turn financial events into explanations that answer questions such as:
What happened?
Why is this worth understanding?
What does it mean in normal language?
What financial concept should I know to understand this?
How could this affect an ordinary person?
I deliberately kept the interface simple: select a date and a region, then get a small set of stories worth understanding.
I also wanted the experience to feel more like discovering an explanation than scrolling through another news app.
Why Does Open Innovation Matter?
Open innovation made it possible for me to experiment with an AI-first idea without building the entire project around a proprietary black-box API.
Using an open-weight model ecosystem meant I could explore different ways of running the model, change the prompts and agent behaviour, and build the application around the model rather than treating AI as a single API call hidden behind the interface.
More importantly, open-source tools made the whole project approachable as a weekend build. The news retrieval, application code, model interaction, and UI can all be inspected and changed.
For a project whose goal is to make complicated information easier to understand, I think that openness matters.
Prize Categories
- Best Use of Render
- Best Use of Gemma
- Best Use of GitHub Copilot
Top comments (0)