MediaFind

MediaFind is a desktop app that indexes your audio and video files entirely on your Mac (or Windows/Linux) and lets you search through them using natural language. Instead of scrubbing through hours of footage, you can type something like "the part where she explains the pricing change" and get ranked results with exact timestamps. It handles transcription, visual scene recognition, on-screen text, faces, logos, music, and more — all without uploading anything to the cloud.

The app works by pointing it at a folder of media files. It then runs a suite of on-device AI models — Whisper for transcription, CLIP for visual search, a local language model for answering questions, and speaker diarization to label who said what. Everything stays on your machine; you can even unplug the internet and it still works. Standout features include a built-in AI assistant that answers questions about your library with cited timestamps, the ability to export search results as FCPXML or EDL timelines for Premiere, Resolve, or Final Cut, and 15 different search modes covering meaning, sound, visuals, emotions, and more.

MediaFind is ideal for anyone who regularly works with recorded media. Podcasters can resurface quotes across hundreds of episodes. Video editors can find b-roll by describing what's on screen. Journalists and researchers can search interviews by meaning and cite exact timestamps. Legal professionals can review depositions privately on-device. Teams can turn meeting recordings into action items. The one-time $29 price (no subscription) and strong privacy focus make it a practical choice for professionals who value both speed and confidentiality.