OpenBrief delivers a desktop app that pulls a video link or local file, downloads it, and runs on‑device transcription and summarization. Built on Tauri v2 with a pnpm/Turborepo stack, it lets users chat over the full transcript or a markdown summary.

The tool uses local models like Whisper for speech‑to‑text and offers a choice of large‑language models via an API key. Summaries appear as timestamped markdown, and users can export notes or play back summaries with text‑to‑speech.

Open source under the GNU Affero GPL, OpenBrief keeps all data on the user’s machine, sidestepping privacy concerns that cloud services raise. The project plans to add support for PDFs, local LLMs such as Gemma 4, and semantic search across a media library.

Developers can clone the repo, run pnpm to install dependencies, and launch the desktop or web app locally. The result is a fully local media‑analysis workflow that blends open‑source tooling with customizable AI models.