https://github.com/raypaste/raypaste
Raypaste is a Mac app built with Tauri that brings fast and customizable AI responses to any app or website.
https://github.com/raypaste/raypaste
developer-tools macos-app native-apps open-source productivity
Last synced: 3 months ago
JSON representation
Raypaste is a Mac app built with Tauri that brings fast and customizable AI responses to any app or website.
- Host: GitHub
- URL: https://github.com/raypaste/raypaste
- Owner: raypaste
- License: mit
- Created: 2026-03-05T07:34:49.000Z (4 months ago)
- Default Branch: main
- Last Pushed: 2026-03-30T19:39:18.000Z (3 months ago)
- Last Synced: 2026-03-30T21:25:47.220Z (3 months ago)
- Topics: developer-tools, macos-app, native-apps, open-source, productivity
- Language: TypeScript
- Homepage: https://raypaste.com
- Size: 1.27 MB
- Stars: 6
- Watchers: 0
- Forks: 0
- Open Issues: 9
-
Metadata Files:
- Readme: README.md
- Contributing: docs/CONTRIBUTING.md
- License: LICENSE.md
Awesome Lists containing this project
README
Raypaste
Fast and flexible AI responses in any app — triggered by a single hotkey.
---
## What is Raypaste?
Raypaste is a native desktop app built with **Tauri 2.0** that brings fast, customizable, and intelligent AI responses to any text input — in any application.
Select text anywhere, on Mac use hotkey `⌘ + Ctrl + R`, and Raypaste transforms it using your configured AI prompts. Results are written back inline or surfaced in a review overlay where you can accept, modify, or decline the changes.
## Platform Support
| Platform | Status | Notes |
| -------- | ---------------------- | ---------------------------------------------------- |
| macOS | ✅ **Fully supported** | Complete functionality |
| Windows | ❌ Not supported | Planned — requires platform-specific implementations |
| Linux | ❌ Not supported | Planned — requires platform-specific implementations |
Raypaste is currently **macOS-only**. While the UI and LLM layer are cross-platform, core integrations (app discovery, focused app detection, selected text capture, and text writeback) rely on macOS-specific APIs.
For detailed compatibility information and contributor guidance on cross-platform support, see **[PLATFORM_COMPATIBILITY.md](docs/PLATFORM_COMPATIBILITY.md)**.
## Features
- **Universal** — works with any app: editors, browsers, terminals, design tools
- **Hotkey-driven** — `⌘ + Ctrl + R` allows you to get an AI response in-place instantly.
- **Per-app prompts** — assign different prompts to different applications for custom repeated workflows
- **Review mode** — preview, stream, and edit AI output in real-time before accepting or rejecting
- **Instant mode** — processing overlay shows live progress while text is being written back directly to your app
- **Cancellation** — press Esc during processing to abort the current request
- **Privacy-first direct mode** — for those wanting to keep data between you and your AI provider only, you can do so with direct-mode. Your prompt and input text goes straight to OpenRouter or Cerebras, never through Raypaste servers, and the output is returned only to your device
- **Prompt library** — create, edit, and manage reusable prompts for your needs
- **History** - view your usage stats and past AI responses (inputs, outputs, prompt used, etc)
## Tech Stack
| Layer | Technology |
| ------------- | ---------------------------------------- |
| Desktop shell | [Tauri 2.0](https://v2.tauri.app) (Rust) |
| Frontend | React 19 + TypeScript |
| Styling | Tailwind CSS v4 |
| State | Zustand |
| LLM providers | OpenRouter, Cerebras, OpenAI |
## Quick Start
```bash
# Install dependencies
pnpm install
# Start dev server + native window
pnpm tauri dev
```
For full setup instructions see the **[getting started guide](docs/DEV_GETTING_STARTED.md)**.
## LLM Providers
Raypaste currently supports three **direct-to-provider** modes (your data never touches Raypaste servers):
- **[OpenRouter](https://openrouter.ai/)** — access to a wide range of models
- [Models used by Raypaste users](https://openrouter.ai/apps?url=https%3A%2F%2Fraypaste.com%2F) — see what models people are using with Raypaste through OpenRouter
- **[Cerebras](https://www.cerebras.ai/)** — ultra-fast inference powered by their Wafer Scale Engine
- [Cerebras Chips](https://www.cerebras.ai/chip) - read more about Cerebras hardware
- **[OpenAI](https://platform.openai.com/)** — direct access to the GPT-5 and GPT-5.4 families in Raypaste
Configure your API keys in the Raypaste app's Settings page (your API keys are stored on your device and do not pass through Raypaste's servers).
The Settings page filters models by provider so users only see models allowed for that direct connection. See **[docs/MODEL_PROVIDER_COMPLIANCE.md](docs/MODEL_PROVIDER_COMPLIANCE.md)** for how provider-scoped filtering and normalization work.
## Contributing
Contributions are welcome — code, docs, bug reports, and feature ideas.
See **[CONTRIBUTING.md](docs/CONTRIBUTING.md)** to get started.
## License
MIT — see [LICENSE](LICENSE).