The core idea behind Vidix is that context switching is the real productivity killer. Every time you leave a document to open a chatbot, re-explain your context and paste the result back, you lose momentum. Vidix collapses that loop into a single hotkey: select the text, invoke a shortcut or the ⌘⇧K palette, and the AI output lands exactly where your cursor already is — either replacing the original, typed in character by character, or opened in an editor for review.
Under the hood, Vidix is built for people who want control over their AI stack. You plug in your own providers — OpenAI, Anthropic, Google Gemini, or Ollama for fully local inference — and decide which model handles which task. There is a workflow canvas for building multi-step agents (a code-review chain with separate security, performance, test-coverage and documentation steps is the canonical example), an image assistant for describing, extracting or generating from images with multimodal models, and a settings panel where shortcuts, prompts and agents are configured without writing code. A status toast reports each stage from capture to paste, and F9 cancels a running action.
Vidix is aimed at Mac users who live inside other apps: developers reviewing code, writers polishing drafts, support and ops people working through Slack and email. It is a native macOS application rather than a web service, and the current release line (0.3.x) ships an Apple-silicon build. Because nothing is stored on Vidix servers and requests go straight to the provider of your choice, the tool is a reasonable fit for teams that are careful about where their text ends up.
