feature-video

To add a video walkthrough to a GitHub pull request, you can automate the capture of browser interactions, encode the screenshots into an MP4 using ffmpeg, and upload the asset to insert the URL directly into the PR body.

Capture a UI demo video for a feature branch by using agent-browser to record the revised flow step-by-step, numbering screenshots to preserve sequence, and then stitching them together with ffmpeg into an MP4 file.

Yes, you need ffmpeg and gh CLI to automate PR video uploads. You also need agent-browser and an authenticated GitHub session to capture screenshots, encode MP4s, validate the user-attachments URL, and insert it into the PR.

You can embed a demo video in a PR description without manual uploads by capturing the localhost walkthrough for the current branch, uploading the generated MP4 via authenticated GitHub sessions, and appending the resulting attachment link to the PR.

The best way to show visual UI changes in a pull request is creating an inline video walkthrough that captures browser interactions, encodes the screenshots into an MP4, and embeds the validated asset URL directly into the PR description for reviewers.

Use agent-browser for recording PR walkthroughs because it captures browser interactions step-by-step while numbering screenshots to preserve the sequence, ensuring the resulting ffmpeg-encoded MP4 accurately demonstrates the revised UI flow.