Local AI-Powered File Search Application for Desktop

This is a fully local, cross-platform AI file search app for macOS and Windows. Instead of hunting for exact filenames, you describe what you're looking for in plain language — "a photo of a mountain at sunset," "the invoice from March," "that Python script with the retry logic" — and the app finds it by meaning, not keyword matching. It embeds and searches images, documents, and code entirely on-device using local ONNX models (CLIP, MiniLM, and a code-specific embedding model) backed by a LanceDB vector store, so nothing ever leaves the user's machine: no cloud calls, no accounts, no telemetry. A Rust engine handles indexing and search, a Go daemon watches the filesystem in real time so the index stays current automatically, and a lightweight native shell wraps a simple web-based UI. Both installers are built and published through an automated CI/CD pipeline.