StockAI is a modern cross-platform desktop application built with Tauri 2.0. It leverages AI technology to perform deep sentiment analysis and scoring on real-time stock news, and provides interactive K-line charts with technical indicators, giving investors data-driven insights in one place.
- Multi-source News Scraping: Automatically collects real-time stock news via Google News RSS (no Chromium needed) with Playwright fallback, supporting US stocks and Chinese A-shares (Shanghai, Shenzhen, and Beijing Stock Exchange).
- Deep AI Analysis: Supports OpenAI (GPT-4o), Anthropic (Claude 3.5 Sonnet), DeepSeek (DeepSeek V4 Pro), GLM (GLM-5.1), and Ollama (local models). Each provider keeps its own API key / base URL / model; switch active provider via a dropdown. Analysis is explicitly triggered โ switching stocks never silently consumes tokens. Deep Mode extracts full article content for richer analysis.
- Interactive K-line Chart: Candlestick chart with MA / BOLL overlays, nine time periods (1D / 5D / 1M / 3M / 6M / YTD / 1Y / 5Y / All), sub-chart indicators (MACD / RSI / KDJ / OBV / VWAP), logarithmic scale, dividend-adjusted price modes, comparison baseline, and real-time price merged into the last candle during trading hours.
- Editable Watchlist: Add and remove stocks freely via the sidebar quick-add input โ the list persists across sessions via local storage.
- Three-Column Dashboard: A dark-theme three-column layout โ left watchlist sidebar, central stock detail panel (real-time price card, K-line chart, news feed), and a right AI analysis panel displaying the sentiment ratio bar, company profile, and color-coded bullish / bearish factor cards. Settings and a global search bar are pinned to the top toolbar.
- Local-first: All API configurations and personalized settings are securely stored locally, never leaving your device.
- Frontend (UI Layer): React 19 + TypeScript + Vite. Responsible for view rendering and user interaction.
- Core Orchestration (Tauri Core): Rust. Manages local persistence, system integration, and Sidecar process scheduling.
- Analysis Engine (Sidecar): Based on the Bun runtime. Uses Playwright for web scraping and integrates AI models for text processing.
Pre-built binaries are available on the Releases page.
macOS Gatekeeper blocks apps that aren't notarized by an Apple Developer certificate. Run this command in Terminal to remove the quarantine flag:
xattr -cr /Applications/StockAI.appThen open the app normally. This is safe โ the app contains no network backdoors and the full source code is auditable in this repository.
Why this happens: Apps downloaded from the internet receive a quarantine attribute. Without an Apple code-signing certificate, macOS shows "damaged" instead of the usual "unknown developer" prompt.
Click More info โ Run anyway. This appears for any unsigned executable.
sudo dpkg -i StockAI_*_amd64.debRequires WebKitGTK (pre-installed on most GNOME-based distros).
- Bun: Primary package manager and Sidecar runtime. Install Bun
- Rust: For building the Tauri core. Install Rust
# Install all dependencies using Bun
bun installTauri's Sidecar mechanism requires specifically named binaries. Compile the Sidecar before running:
# macOS ARM64 (Apple Silicon)
bun build sidecar/index.ts --compile --outfile sidecar/stockai-backend-aarch64-apple-darwin
# Windows x86_64
bun build sidecar/index.ts --compile --outfile sidecar/stockai-backend-x86_64-pc-windows-msvc.exe
# Linux x86_64
bun build sidecar/index.ts --compile --outfile sidecar/stockai-backend-x86_64-unknown-linux-gnubun tauri devRun all tests with a single command (no GNU timeout required):
bun run testIndividual layers:
- Frontend Tests (Vitest):
bunx vitest run - Sidecar Logic Tests (Bun):
cd sidecar && bun test - Rust Core Tests (Cargo):
cd src-tauri && cargo test - Integrated Smoke Test:
bun scripts/smoke-test.ts
- Desktop Framework: Tauri 2.0 (Rust)
- Frontend Framework: React 19, TailwindCSS 4, Lucide Icons, Lightweight Charts
- Scraper/Backend: Bun, Playwright, NodeHtmlMarkdown
- AI Integration: OpenAI SDK, Anthropic SDK, Ollama SDK (DeepSeek / GLM via OpenAI-compatible API)
- Code Comments: All logic comments must use Chinese (as per project preference).
- Architecture Principles: Strictly follow Clean Architecture with unidirectional dependency flow (UI -> Core -> Sidecar).
- Test Driven: All parsing logic must be verified by offline unit tests.
MIT License