Promptify 🚀

Promptify is a sleek, premium, high-fidelity prompt engineering dashboard built to optimize and generate structured, token-light prompts. It is designed to act as a "drafting assistant" to optimize raw thoughts, logs, and requirements before pasting them into heavier online LLM models.

Promptify supports local models (via Ollama) as well as cloud-hosted APIs (Google Gemini, OpenAI, Anthropic Claude, OpenRouter, and DeepSeek).

---

✨ Features

1. Dual Workspace Modes:

   • ⚡ Prompt Optimizer: Paste raw, unstructured notes, event logs, or code, select a template preset, and get a compacted, high-density prompt.
   • 🛠️ Prompt Generator: A guided, step-by-step form to build custom prompts from scratch by defining the Core Task, Persona, Context, Target Audience, Tone, Deliverable Format, Constraints, and Few-Shot Examples.

2. 6 Model Providers & Real-Time Streaming:

   • Ollama: Local models running on your machine (e.g. Qwen, Gemma, Llama).
   • Google Gemini: Access to Gemini 2.5 Flash and Pro.
   • OpenAI: Access to GPT-4o-mini, GPT-4o, and o3-mini.
   • Anthropic Claude: Access to Claude 3.5 Sonnet and Claude 3.5 Haiku.
   • OpenRouter: Access to Llama 3.3, Gemini, Claude, and DeepSeek.
   • DeepSeek: Access to DeepSeek V3 and DeepSeek R1 (Reasoner).
   • Token-by-token streaming is fully supported across all 6 providers.

3. Desktop Native (Tauri) & Mobile Responsive:

   • Built-in support to run as a native desktop application on macOS, Windows, and Linux via Tauri.
   • Fully responsive CSS layout: automatically collapses to a stacked column layout on mobile screens (<768px).

4. Productivity Tools:

   • Visual Diff engine: Side-by-side view showing removed words (in red strikethrough) and additions (in green), helping you track exact token compression.
   • Auto-Copy Output: Instantly copies optimized prompts to your clipboard as they stream in.
   • Local Storage Sync: Automatically persists history logs, custom rules, and API settings in your local browser storage.
   • Auto-Draft Meta-Preset: Propose a task, and let Promptify auto-write a custom system prompt template.

---

🛠️ Getting Started

1. Installation

Clone the repository and install the dependencies:

npm install

2. Run the Web App

Start the Vite development server:

npm run dev

Open your browser to http://localhost:5173.

3. Run as a Desktop App (Tauri)

To launch the native desktop application in development mode:

npm run desktop:dev

To compile a standalone desktop installer (.app on macOS, .exe on Windows):

npm run desktop:build

---

🔌 Connecting to Local Ollama Models (CORS & PNA Setup)

Web browsers block requests from public websites (like a Vercel deployment) to local addresses (localhost) due to CORS and Private Network Access security policies.

To allow Promptify to query your local Ollama models, you need to enable the OLLAMA_ORIGINS variable. Here is how to configure it permanently so you don't have to run it from a terminal every time:

1. macOS (Mac GUI App)

If you run Ollama as a native macOS Menu Bar application:

1. Open your terminal and append the environment declaration to your profile:

   echo 'launchctl setenv OLLAMA_ORIGINS "*"' >> ~/.zprofile

2. Apply the setting for the current session:

   launchctl setenv OLLAMA_ORIGINS "*"

3. Restart the Ollama App: Quit Ollama from the macOS Menu Bar and re-open it.

---

2. Windows (Tray App)

If you run Ollama as a Windows background task:

1. Quit Ollama: Close the Ollama application by right-clicking the taskbar tray icon.
2. Set Environment Variable:
   • Open the Windows Start Menu, search for "Environment Variables", and select Edit environment variables for your account.
   • Under User variables, click New....
   • Set Variable name to OLLAMA_ORIGINS and Variable value to *.
   • Click OK to save.
3. Launch Ollama: Open Ollama again from the Start Menu.

---

3. Linux (Systemd Service)

If you installed Ollama as a service:

1. Open the systemd override config editor:

   sudo systemctl edit ollama.service

2. Add the environment rule in the file:

   [Service]
   Environment="OLLAMA_ORIGINS=*"

3. Save the file and reload systemd to apply:

   sudo systemctl daemon-reload
   sudo systemctl restart ollama

---

💡 Best Alternative: Use the Tauri Desktop App

If you do not want to configure CORS on your system, simply use our native Desktop Application (via npm run desktop:dev or building the binary). Because native desktop apps run outside of the web browser's security sandbox, they are fully exempt from CORS and Private Network Access blocks and will connect to your local Ollama server out of the box!

---

📂 Project Structure

├── public/                 # Static vector SVG assets and favicon
├── src/
│   ├── services/
│   │   └── api.js          # API clients for all 6 providers (Ollama, Gemini, OpenAI, Claude, etc.)
│   ├── App.jsx             # Main dashboard layout, states, diffs, and history logs
│   ├── index.css           # Glassmorphism dark styles & mobile media queries
│   └── main.jsx
├── src-tauri/              # Tauri native desktop application configs
├── package.json
└── vite.config.js

📄 License

This project is licensed under the MIT License.