A short explanation is useful. A searchable transcript is better.
Download the video
Save the best available MP4 from a supported X post or Google Drive preview. The extension works with primary, quoted, retweeted, and nested X videos.
Local video workflow for X and Google Drive
A free Chrome extension that adds Download and Transcript actions directly to video posts. Keep the MP4, copy the words, or export clean TXT and SRT files — without sending audio to our servers.
Chrome-compatible browser · macOS · Apple Silicon · Spokenly
A short explanation is useful. A searchable transcript is better.
Built for researchers, writers, editors, developers, and anyone who would rather search, quote, summarize, or archive a video than replay it five times.
Save the best available MP4 from a supported X post or Google Drive preview. The extension works with primary, quoted, retweeted, and nested X videos.
FFmpeg prepares temporary audio, then Spokenly runs NVIDIA Parakeet TDT 0.6B v3 locally through its Apple-optimized Core ML pipeline.
Copy clean text, add timestamps every 10, 30, or 60 seconds, download TXT, or create sentence-level SRT subtitles.
Completed transcripts stay in browser-local history for your chosen retention window. Reopening the same video can reuse the saved result instead of rerunning the model.
The extension is currently installed from GitHub. A small Native Messaging host connects the browser to local FFmpeg and Spokenly — no hosted transcription account required.
Install FFmpeg and Python, then install the current sideload macOS build of Spokenly.
Select NVIDIA Parakeet TDT 0.6B v3 for file transcription and enable Spokenly's local MCP server.
Clone this repository, turn on Developer mode, and choose Load unpacked.
Run the included installer once, reload the extension, and open a supported video.
$ brew install ffmpeg python $ git clone \ https://github.com/Lemelson/x-post-toolkit.git $ cd x-post-toolkit $ ./scripts/install_native_host.sh Native host installed.
Audio stays local. Your transcript stays yours.
Free and open source
Install the developer release from GitHub and use the model already managed by Spokenly.