Distill News is an automated news website made using multi-agents with the simple goal in mind - to surf the internet and find every possible news.
News articles written by media channels can not cover everything due to a bias towards profit and limited workforce.
While the internet is packed with all sorts of news all over the place on different platforms with no unified platform where one can find all the news, something might be posted on Reddit or X, it is impossible to follow the fast-paced nature of the internet.
This is where our goal comes in - using AI agents to find news from data all over the web from news articles to community posts and present them in a structured, personalized, and timely manner for engagement.
An AI-driven platform that cuts through the noise, bias, and bloat of traditional media by extracting core news from web scraping, trusted APIs, and real-time community sources like Reddit.
The system strips clickbait and filler content found in news articles, and as well as rewrites news and community posts into clear, professional-grade articles using Julep AI.
By combining structured extraction (via Julep AI), community-driven discovery, and smart personalization, we empower communities with direct access to reliable local and global updates—without clickbait, corporate spin, or media gatekeeping.
This approach also enables underrepresented voices on social platforms to reach a wider audience through well-written, structured articles.
AI-Generated News Feed: Aggregates content from APIs (GNews, MediaStack, RapidAPI) and community platforms like Reddit.
Noise Filtering: Removes clickbait, ads, duplicates, and irrelevant content automatically.
Multi-Agent Article Processing (via Julep AI):
Classifies input data as news or not. Extracts structured content (title, summary, date, location). Formats it into clean, professional markdown
Personalized Feed: Learns from your reading time and interests. Dynamically adapts story selection using topic-weighted preferences
RAG-based AI Assistant: Includes article context automatically. Supports session memory for follow-ups. Handles both local and global queries conversationally.
Clean UI: Minimalist, responsive interface with login and onboarding flow.
- Julep AI
- Next.js
- FastAPI
- BeautifulSoup
- nodemailer (javascript)
- Python
- Typescript
- Javascrip
- YAML
We want to build an AI-driven platform to:
- Aggregate core news from web scraping, trusted APIs, and community sources.
- Filter out filler content, ads, and fluff.
- Rewrite articles and community posts into concise, professional-grade news.
- Personalize delivery based on interests, reading behavior, and location.
-
Concise Rewriting
- Removes filler, ads, and fluff.
- Rewrites content for clarity and brevity.
-
Reddit & Local Feeds
- Surfaces hyperlocal stories directly from community posts and public forums.
-
AI-Enhanced Writing
- Converts raw user-generated content into professional-quality news articles using Julep AI.
-
Personalized Recommendations
- Delivers news based on user interests, reading behavior, and location.
-
Curated Newsletters
- Sends daily or weekly digests tailored to each subscriber.
-
Print-Style Digital Newspaper
- Offers a traditional-style newspaper layout with various sections as an alternative to typical email newsletters.
-
Human Validation (Optional)
- Allows moderators or volunteers to fact-check and approve stories when needed.
Built with Julep AI and open-source community data.
cd frontend
npm installpip install -r requirements.txt
fastapi dev server/app.pycd frontend
npm run devMake sure to add .env file and API keys
Team Name: SNAP_back_to_reality