Buza Studio is a local-first prompt engineering environment designed for developers and AI engineers. It offers a distraction-free workspace for crafting, testing, and optimizing LLM prompts with complete data privacy.
Just like Obsidian changed how we take notes, Buza Studio aims to change how we engineer prompts:
- Local-First & Private: Your prompts live on your machine. No cloud lock-in, no data telemetry.
- Distraction-Free: A clean, minimal interface that lets you focus on the text.
- Fast & Native: Built with Tauri and Rust for blazing fast performance.
- Your Data, Your Rules: You own your prompt library. Version it with Git, sync it how you want.
- ⚡ Local-First & Fast: Built on Tauri and Bun for native performance. Your prompts stay on your machine, ensuring complete data privacy.
- 🧠 Smart Variable Tracking: Automatically detects and tracks variables in your prompts using
{{variable}}syntax. - 🤖 AI Integration: Built-in support for Google Gemini to generate prompt structures, optimize content, and run variations.
- 🔄 Version Control: Save versions, restore history, and manage prompt variants effortlessly.
- 🧩 Template Library: Start quickly with built-in templates or save your own for reuse.
- 📝 Distraction-Free Editor: A clean, focused interface designed for writing and iterating on complex prompts.
- Core: Tauri v2 (Rust + Webview)
- Frontend: React 19, Vite
- Styling: TailwindCSS v4, Shadcn UI
- State Management: Jotai
- Runtime: Bun
- Bun: Install Bun (
curl -fsSL https://bun.sh/install | bash) - Rust: Install Rust (Required for Tauri)
- Node.js: (Optional, but recommended for some tools)
-
Clone the repository
git clone https://github.com/yourusername/buza-studio.git cd buza-studio -
Install dependencies
bun install
-
Set up Environment Create a
.envfile in the root directory and add your API keys:VITE_GEMINI_API_KEY=your_api_key_here
Start the development server with Tauri:
bun tauri:devThis will launch the native application window.
Contributions are welcome! Please feel free to submit a Pull Request.
MIT