Case studies — WEB · CLOSED 2024
60+ sources. Real-time sentiment. A terminal ahead of its time.
BlockBeat aggregated crypto news across 60+ sources with market sentiment scoring. Shut down in 2024.
BlockBeat ran for 18 months as a real-time crypto intelligence terminal, a cyberpunk-aesthetic newsroom where every article arrived with a sentiment score, every coin had a rolling sentiment curve, and an AI newsletter hit inboxes every morning summarizing the previous 24 hours. Sunset in 2024.
A full crypto intelligence stack.
Real-time feed with sub-minute latency from publication. Every major crypto media outlet, financial wire, and influential blog, centralized in one terminal.
Each article scored bullish / bearish / neutral using NLP. Scores aggregated per coin (BTC, ETH, SOL, and more) and displayed as rolling sentiment curves updated in real time.
Live coin prices surfaced alongside relevant news, TradingView widgets embedded inline, Supabase-cached price data updated on a tight polling interval.
Every morning an AI-written newsletter hit subscribers' inboxes, a concise summary of the previous 24 hours in crypto, automatically generated from the aggregated feed.
A terminal, not a blog.
Full-bleed dark UI. Monospace type throughout. Green-on-black data displays. Dense information layout designed for power users.
Filter by publication, coin mention, sentiment polarity, and time range. Build a custom feed with a few clicks.
BlockBeat ran for a year and a half before being shut down in 2024. The architecture lives on, the real-time aggregation and NLP pipeline have been adapted for other projects.
What it ran on.
On the bench.
- Next.js
- Supabase
- NewsAPI.ai
- TradingView
- NLP pipeline