ModelFlux is an open-source, cross-platform AI chat application that lets you interact with various Large Language Models. You can use cloud providers like OpenAI and Anthropic, connect to local servers like Ollama, or run models directly on your device using ExecuTorch or Llama.cpp.

Yes, ModelFlux is completely free and open source. You only pay for the API usage if you use cloud providers like OpenAI or Anthropic directly. Local models and Ollama are completely free.

Absolutely! On Android devices, you can download and run local models using ExecuTorch (.pte) or Llama.cpp (.gguf). Once downloaded, these models work completely offline with no internet required.

ModelFlux supports Android, and Web platforms. The native Android app has full support including local on-device models. The Web version supports remote providers (OpenAI, Anthropic, Ollama). iOS support is coming soon.

Yes! All your conversations and settings are stored locally on your device. We don't collect any analytics or usage data. When using cloud providers, your data goes directly to their APIs - we never store or intercept it.

RAG (Retrieval-Augmented Generation) lets you add documents (PDFs, text files) to your conversations. The app extracts and embeds the content, then uses it to provide context-aware responses. This works with both local and cloud models.

Personas are custom AI personalities you can create with system prompts. Based on the Character Card V2 specification, they let you customize how the AI responds - from helpful assistants to specialized experts or creative characters.

Go to Settings > LLM Providers > Add Provider. Choose from OpenAI, Anthropic, Ollama, or OpenAI-Compatible (for services like LM Studio, Together AI, etc.). Enter your API key and base URL as needed.

Yes! You can change the LLM provider or model at any point during a conversation. Previous messages are preserved, and future responses will use the newly selected model.

Go to Settings > Data Management > Export Data. Your conversations, personas, and settings will be exported as a JSON file that you can import later or use as a backup.