witsy

<div align="center">
  <a href="https://witsyai.com" target="_blank"><img src="assets/icon.png" width="128" alt="Witsy Logo"></a>
  <div><b>Witsy</b></div>
  <div>Desktop AI Assistant</div>
</div>

<div align="center">
Witsy is a desktop AI assistant that supports multiple LLM providers through API keys (BYOK) or local models via Ollama. It offers features like chat completion with vision models, text-to-image/video generation (OpenAI, Google, xAI, Replicate, fal.ai, HuggingFace), and image-to-image editing.
</div>

Downloads are available at [witsyai.com](https://witsyai.com) and the [releases](https://github.com/nbonamy/witsy/releases) page.

Key functionalities include a scratchpad for content creation, "Prompt Anywhere" for generating text in any application, and AI commands for text manipulation. It also features expert prompts, LLM plugins (Python code execution, internet search), long-term memory, text-to-speech, and RAG for chatting with local files.

Witsy supports providers like OpenAI, Ollama, Anthropic, MistralAI, Google, xAI, OpenRouter, DeepSeek, and Groq. Setup involves installing the application and configuring API keys for desired services.
</div>