Freechat.run
FreeChat is an AI-powered chat application designed specifically for macOS. It provides users with a local and secure platform for communic...
Last verified:
What is Freechat.run?
FreeChat is a local, secure, open-source AI chat application designed specifically for macOS. It allows users to chat with large language models (LLMs) completely offline on their Mac without installing any other software or configuring complex settings. Every conversation is saved locally on the device, ensuring complete privacy with no data sent to external servers or conversation tracking.
Key features include the ability to customize persona and expertise by changing the system prompt, support for any GGUF-formatted model compatible with llama.cpp (including CodeLlama, Mistral 7B, Llama 2, and Llama 3), and no internet connection required after installation. The app is built using the open-source llama.cpp library and runs as a native macOS application with a simple interface similar to Apple's Notes app.
FreeChat is ideal for macOS users who want privacy-focused AI chat, developers who want to audit their AI software, privacy-conscious individuals who don't want conversations sent to cloud servers, and non-technical users who want to experiment with local AI without complex setup. The app is free on the Mac App Store with no subscriptions.
The application prioritizes performance and simplicity over features, following a
Freechat.run pricing
Pricing model: Free
Free - The app is completely free on the Mac App Store with no subscriptions, no paid tiers, and no in-app purchases. There are no premium features or paid plans. The default model is included with the 3GB app download, and users can download additional GGUF models for free from Hugging Face. The developer explicitly states 'No subscriptions, no snooping, complete freedom.'
Freechat.run pros
- Completely offline operation - no internet required after installation
- All conversations saved locally with no data collection
- Open source and auditable code for transparency
- No subscriptions, no login, no tracking
- Compatible with any GGUF-formatted llama.cpp model
- Supports CodeLlama, Mistral 7B, Llama 2, and Llama 3
- Customizable system prompt for persona and expertise
- Simple native macOS interface - no technical setup needed
- Free on Mac App Store with no paid tiers
- Active development with regular llama.cpp updates
- Privacy-focused - talk about anything like using Notes
- Small app size (21.3 MB)
- Works beautifully on Apple Silicon (M1, M2 chips)
- No conversation tracking or external server reliance
- Downloadable from Mac App Store for instant use
Freechat.run cons
- MacOS only - not available for Windows, Linux, or mobile
- Requires 8GB+ RAM to run models
- Requires macOS 13.0 (Ventura) or later
- Only 21.3 MB app but models can be large (default 3GB)
- No built-in model download feature yet (coming in roadmap)
- More quantization makes models less intelligent
- Larger parameter models run slower
- Some models like DeepSeek may cause freezing issues
- No advanced settings exposure yet (temperature, prompt format)
- No conversation search feature yet (on roadmap)
Frequently asked questions about Freechat.run
What is FreeChat?
FreeChat is a native macOS application that lets you chat with large language models (LLMs) completely offline on your Mac. It's built on llama.cpp and requires no configuration - just download from the Mac App Store and start chatting. Every conversation is saved locally with no data sent to external servers.
Do I need an internet connection to use FreeChat?
No internet connection is required after installation. FreeChat runs completely locally on your device. The only time you need internet is when downloading the app from the Mac App Store and when downloading additional GGUF models from Hugging Face.
Is FreeChat really private and secure?
Yes. Every conversation is completely private and saved only on your local Mac. The developer does not collect any data from the app. There is no conversation tracking, no login required, and no data sent to external servers - just like using Apple's Notes app.
What AI models does FreeChat support?
FreeChat is compatible with any GGUF-formatted model that works with llama.cpp. This includes CodeLlama, Mistral 7B, Llama 2, Llama 3, and many others. Models are typically named with their parameter count (e.g., 7B, 13B) and come in different quantization levels.
How much RAM do I need to run FreeChat?
FreeChat requires 8GB+ RAM to run AI models. Larger models with more parameters will need more RAM and run slower. The exact RAM requirement depends on which model you choose to load.
Can I customize the AI's personality or expertise?
Yes. You can customize the persona and expertise by changing the system prompt in the app. The roadmap also includes a 'Personas' feature that will let you save system prompts and model settings for later use when creating new conversations.
Where can I download additional AI models?
You can find GGUF-formatted models on Hugging Face. Most models have a 'model card' by the author that discusses their training and abilities. The AI training community releases new models regularly.
Is FreeChat open source?
Yes, FreeChat is completely open source and hosted on GitHub. This allows anyone to audit the code to verify that it actually runs locally and doesn't send data externally. The source code is available at github.com/psugihara/FreeChat.
What macOS versions does FreeChat support?
FreeChat requires macOS 13.0 (Ventura) or later. It works on Mac computers including MacBook Air, MacBook Pro, and iMac with Apple Silicon (M1, M2 chips) or Intel processors that meet the RAM requirement.
Are there any paid features or subscriptions?
No. FreeChat is completely free with no subscriptions, no paid tiers, and no in-app purchases. The app is free on the Mac App Store, and all features are available without any payment. The developer explicitly states 'No subscriptions, no snooping, complete freedom.'