https://github.com/rinpatch/spearline
https://github.com/rinpatch/spearline
Last synced: 11 months ago
JSON representation
- Host: GitHub
- URL: https://github.com/rinpatch/spearline
- Owner: rinpatch
- Created: 2025-06-20T12:11:59.000Z (about 1 year ago)
- Default Branch: master
- Last Pushed: 2025-06-21T20:43:10.000Z (about 1 year ago)
- Last Synced: 2025-06-21T21:30:42.684Z (about 1 year ago)
- Language: TypeScript
- Homepage: https://imaginehack-2025.vercel.app
- Size: 1.03 MB
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Spearline - Piercing Bias. Truth in the Line.
Spearline is an AI-powered Malaysian news aggregator designed to uncover media bias.
We bring together news from across languages — Malay, English, and Chinese — and use AI to summarize each article and identify its ideological leanings, such as Pro-Government, Pro-Islam, or Opposition-aligned.
## Team
This project is done by Nott-A-Hackers team for Taylor's Univeristy ImagineHack 2025 Hackathon. The team members are:
- [@rinpatch](https://github.com/rinpatch)
- [@Mint3Ds](https://github.com/Mint3Ds)
- [mithileshtew0702](https://github.com/mithileshtew0702)
- [@Yoonjae7](https://github.com/Yoonjae7)
- [Pyjalal](https://github.com/Pyjalal)
## Technologies Used
- Next.js (FE/BE)
- Vercel (Deploy)
- Supabase (Postgres Database to store articles/embeddings)
- Qstash (Job queue for news scraping/processing)
- Upstash Redis (In-Memory cache, we just use it to keep track of recently processed articles to not reprocess them)
- Openrouter (Unified LLM API allowing for easy switching of models, we currently use gemini-2.0-flash for everything due to good cost-to-benefit)
- OpenAI embeddings API (used to transform news text into number vectors, which we can compare for similarity in order to "cluster" articles about same topic into one story)
## Deploying
Deploy in a [Vercel](https://vercel.com) project with QStash, Supabase and Upstash Redis connected. Set the following evironment variables:
- `APP_BASE_URL`: prod url that QStash cron will query (i.e https://spearline.vercel.app/).
- `OPENROUTER_API_KEY`: OpenRouter API key to be used for analysis as well as summary/title generation.
- `OPENAI_API_KEY`: OpenAI API key, will only be used for text embeddings.
## Challenge and Approach
### Problem
Malaysia's news media is deeply fragmented along ethnic, linguistic, and political lines. A single news story can be interpreted in completely different ways depending on whether the reader views it in Malay, Chinese, or English — and from which outlet.
### Our Approach
Spearline tackles this problem by using AI to detect and visualize ideological bias in news articles across multiple Malaysian media sources. We built an aggregator platform that:
- Collects news articles from different language-based sources
- Uses an LLM (via OpenRouter) to summarize articles and classify bias (e.g., Pro-Government, Pro-Opposition, Pro-Malay, etc.)
- Displays these biases in a clean, percentage-based breakdown per article
- Offers users different perspectives side by side, helping them see beyond their media bubble
**The goal**: Make media bias transparent, and empower Malaysians to think critically about what they read.
## Features
- **AI-Powered Bias Detection**: Uses advanced LLM technology to classify ideological leanings
- **Visual Bias Breakdown**: Clean percentage-based visualization of detected biases
- **Side-by-Side Comparison**: Compare different perspectives on the same story
- **Real-time Analysis**: Live processing of news articles as they're published
## Usage Instructions
Visit: [https://spearline.vercel.app]
## AI Usage
In accordance with hackathon rules, we have utilized Cursor editor for help with most modules of this project.