A cross-platform LLM gateway and debugging workbench — focused on protocol routing, model debugging, security probing, with provider/key management built in.
中文说明请见
README.zh-CN.md。Current version:
2.0.0.
- Provider management: Add, edit, and delete AI providers (e.g. OpenAI, Anthropic)
- Key management: Add multiple API keys per provider
- Expiry tracking: Set expiration dates and automatically mark expired keys
- Quick copy: One-click copy provider base URL and API keys
- Model search: Search models by name or ID in the model list
- Model testing: Test individual models to verify availability (supports OpenAI & Claude protocols)
- LLM gateway (Anthropic-compatible): Route Claude model IDs to configured providers/models with a gateway token
- Gateway runtime control: Start/stop gateway process and inspect runtime logs in-app
- Local storage: All data stays on device, nothing is uploaded
- Cross-platform: Windows, macOS, and Linux
- Lightweight: Built with Tauri, around 10MB
Five key pages are shown below.
Provider and key overview, including key status and quick actions.
Search and review available models under the selected key/provider.
Run single debug requests with model/protocol controls and inspect responses.
Execute multi-case security probing and view pass/warn/fail matrix results.
Configure model routing and manage gateway runtime status/logs.
- Node.js 18+
- npm
- Rust 1.88+ (for building Tauri)
# macOS
curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh
# Windows
# Download and run https://rustup.rs/
# Linux
curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | shnpm install# Start Tauri dev mode (will also start Vite)
npm run tauri:dev# Build for current platform
npm run tauri:build
# Debug build (faster)
npm run tauri:build:debugBuild outputs are generated at src-tauri/target/release/bundle/.
Automated releases only produce macOS and Windows binaries, so Linux users must build locally if they need a native release.
- Install the system dependencies listed under
src-tauri/tauri.conf.jsonand the workflow (glib, GTK, WebKit2GTK, etc.). - Run
npm cifollowed bynpm run tauri:build. - Grab the bundle from
src-tauri/target/release/bundle/.
src/
app/ App entry and layout
domains/ Domain modules
providers/ Provider management
keys/ API Key management
settings/ Security and settings
shared/ Shared components and utilities
i18n/ Localization resources
types/ Global type definitions
- Click the "Add Provider" button in the left sidebar
- Enter the provider name (e.g. "OpenAI")
- Enter the provider base URL (e.g. "https://api.openai.com")
- Select a provider
- Click the "Add Key" button in the top-right corner
- Enter the API key (required)
- Optional: add a name, note, and expiration date
- 🟢 Green: Key is valid
- 🟡 Yellow: Key expires soon (within 7 days)
- 🔴 Red: Key has expired
- Tauri 2.0: Lightweight cross-platform desktop framework
- React: UI framework
- Vite: Build tool
- Tailwind CSS: Styling
- Zustand: State management
- TypeScript: Type safety
- Rust: Backend logic
The app uses a Soft UI (New Soft UI) visual style featuring:
- Soft shadows and rounded corners
- Purple-themed palette
- Frosted glass effects
- Smooth transitions
- Strong accessibility
Compared to Electron, Tauri offers:
- Smaller size: ~10MB vs Electron’s 150MB+
- Lower memory usage: Uses the system WebView
- Better security: Rust memory safety
- Faster startup: Lighter runtime
This project was fully developed with Codex and Claude Code.
Special thanks to both tooling teams for enabling rapid iteration from product ideas to implementation.
MIT License. See LICENSE.
Issues and PRs are welcome. Please read CONTRIBUTING.md first.