What is dailybrief?

DailyBrief is a self-hosted news aggregator that pulls content from 23 sources, runs it through an AI language model, and produces a single daily briefing in Chinese or English. The sources span GitHub's trending projects, AI research blogs (OpenAI, DeepMind, Hugging Face, MIT Technology Review), major international news outlets (BBC, Guardian, New York Times, NPR, Al Jazeera), financial publications (Bloomberg, Wall Street Journal, Financial Times), and community forums like Hacker News or Chinese tech communities. None of these sources require a paid API key: everything uses publicly available RSS feeds and open JSON endpoints. Alongside the news summary, the tool tracks 21 financial assets in real time: US stocks and ETFs (SPY, QQQ, NVDA, TSLA, and others), cryptocurrencies (BTC, ETH, SOL), Hong Kong and Chinese equities, gold, crude oil, currency pairs, and macro signals like the VIX volatility index and 10-year Treasury yield. For each asset it computes technical indicators (SMA, RSI, MACD) and asks the AI to write a short daily commentary on the price action. You can run the whole thing in three ways. The first is a GitHub Actions setup where you fork the repository, add your API key as a secret, and GitHub runs the workflow on a schedule, publishing the resulting single-file HTML report to GitHub Pages. No server is needed and the cost of AI API calls is roughly a dollar or two per month. The second way is a one-line local install script that clones the project, sets up a system scheduler, and wires everything up automatically on macOS, Linux, or Windows. The third way is to paste a short instruction to an AI coding agent like Cursor or Claude Code and have it handle the setup. The output is a single self-contained HTML file with all CSS and JavaScript inlined, so it needs no web server or external dependencies to display. The tool supports five AI backends (Anthropic Claude, OpenAI, DeepSeek, MiniMax, or any OpenAI-compatible proxy) and lets you switch between them with one environment variable. The full README is longer than what was shown.