I built Wisp because I kept alt-tabbing between my editor/browser/docs and chat apps, copying context over, waiting for an answer, then copying the result back. Wisp is a desktop overlay for using AI from whatever app you are already in. -You can prompt with your choice of context (Text, screenshot, snipping, active app etc) in a few key strokes -For longer back-and-forth, it can open into a full chat window. -Supports hold-to-talk voice queries, dictation into any field, local TTS, -Can pastes the result back directly over selected text. -Data is local first, and nothing is sent to provider without your permission -It supports 20+ local and remote providers, including OpenRouter, Groq, Gemini, Ollama, and LM Studio. -Runs on Windows, macOS and Linux v0.9 Added Wisp Context Server: trusted MCP clients can read selected text, clipboard, active window, browser-page text, and screen snippets. I no longer have access to Mac hardware, so I would especially appreciate macOS testing. Bug reports and ideas are appreciated. Repo + demo GIFs: https://github.com/SunnyLich/Wisp-AI-Assistant Website: https://sunnylich.github.io/Wisp-AI-Assistant/#overview