Craft Perfect AI Prompts
Completely Offline.
Free Forever
Your Private
Built with Electron and the T3 stack, PromptCrafter brings state-of-the-art AI to your desktop or laptop without compromising your privacy.
100% Private & Offline
Your prompts and data never leave your computer. All processing happens locally, giving you complete control and confidentiality.
Powered by Google Gemma 3
Leverage the efficiency of Google's Gemma 3 models. Choose from various sizes available in Ollama: 1B, 4B, 12B, or 27B parameters to match your hardware and needs.
Simple & Accessible
Runs on any modern computer with at least 8GB of RAM and a modern CPU. A super simple setup UI gets you connected to your local Ollama instance in seconds.
Ollama is Required for Local Use
To run the Gemma 3 models on your own computer, you need to have Ollama installed. Ollama is a fantastic tool that makes it incredibly simple to run powerful large language models locally. Through Ollama, you just need to download your preferred Gemma 3 model, which is offered completely free of charge. Just make sure Ollama stays running in the background while using PromptCrafter.
Local Models & Hardware
Choose the right Gemma 3 model for your needs and learn what it takes to run it locally with Ollama.
Hardware Recommendations
1B Parameter Model
Recommended RAM: 8 GB+
Ideal for basic tasks and older hardware. Runs well on most modern CPUs.
4B Parameter Model
Recommended RAM: 16 GB+
A great balance of performance and resource usage. A dedicated GPU is recommended for better speed.
12B Parameter Model
Recommended RAM: 24 GB+
For more complex tasks. Requires a powerful CPU and a modern dedicated GPU for smooth operation.
27B Parameter Model
Recommended RAM: 48 GB+
The most powerful model. A high-end GPU (NVIDIA RTX) is strongly recommended.
About the Gemma 3 Models
- Built with the same research and technology as Google's powerful Gemini models.
- The 4B, 12B, and 27B models are multimodal, capable of understanding both text and images.
- Features a massive 128,000 token context window (for 4B+ models), allowing for analysis of very large documents.
- Trained with support for over 140 languages, making it a truly global model family.
- Includes official Quantization-Aware Trained (QAT) versions, which reduce size and computational needs while maintaining high accuracy.
Download PromptCrafter
Stay tuned for the official PromptCrafter release on Windows, Mac, and Linux!
Windows
macOS
Linux
When will PromptCrafter be released?
Official realase target is early-mid November 2025. Stay tuned!
If you would like to be contributor, visit the open source GitHub repository!️
Proudly Open Source
PromptCrafter is built by AI enthusiasts for the AI community. Check out the source code, report issues, or contribute to the project on GitHub.
View on GitHub