Chatty-Mini

Chatty-mini: Your Personal, Private AI Workspace

Chatty-mini gives you a real generative AI assistant that can run directly on your Android device — even without an internet connection.

It is completely free, with no ads and no in-app purchases. A local AI model is included with the app, so you can install Chatty-mini and start chatting straight away without hunting for models, creating an account, or relying on a cloud service.

AI that works offline

Chatty-mini includes a lightweight on-device AI model designed for everyday conversation, writing, brainstorming, notes and experimentation.

Once installed, local AI features can work without an internet connection. Your local chats and files stay on your device by default.

Bring your own models too

Want more control? Chatty-mini also supports compatible GGUF models.

You can import your own GGUF files and assign different models to roles such as Main AI and Bookkeeper, allowing you to experiment with different sizes, personalities and capabilities.

The included model is there to make Chatty-mini useful immediately. Bringing your own model is optional.

Built for small Android devices

Chatty-mini uses a compact, portrait-first interface designed to make local AI practical on phones rather than squeezing a desktop-style workspace onto a small screen.

The app includes:

• Built-in local AI that works offline
• Optional support for your own GGUF models
• Main AI and Bookkeeper model roles
• Local project sandbox for .md, .txt and .json files
• Saved character and persona profiles
• No ads
• No in-app purchases
• No mandatory cloud connection

Optional cloud AI

Cloud AI is available only if you choose to use it.

Chatty-mini can connect to supported services such as OpenAI, Anthropic, Gemini, xAI and DeepSeek using your own API credentials. API keys are stored locally using Android security features, and cloud requests happen only when you direct the app to use them.

There is no hidden automatic cloud fallback.

Privacy first

Local conversations and files are not automatically sent to an external server.

When using the built-in local AI or another local GGUF model, inference happens on your device. If you deliberately choose a cloud provider, the information required for that request is sent to the provider you selected.

Free and open source

Chatty-mini is free and open-source software released under the GNU Affero General Public License v3.0 (AGPLv3).

You can inspect the source code, modify it, redistribute it and see how the software works.

AI disclosure

Chatty-mini provides generative AI capabilities. AI-generated responses can be incomplete or incorrect, so important information should always be independently verified.

Recent changes:
Bundled SmolLM2-360M model for out-of-the-box local chat
Fixed FMI contact bug