A Product Hunt gallery is a scroll, and a launch video's first job is to survive that scroll. Most visitors are moving past dozens of thumbnails, many of them autoplaying muted, before deciding which one is worth a real click. A video built for this moment has to earn attention in its opening frame and its first few seconds, well before a viewer commits to watching the whole thing or turning sound on. That is a different constraint than a landing page video, which only has to satisfy a visitor who already chose to be on that page.
For a Bolt built app, the practical starting point is the same question that applies before any other public facing asset: does the app have a stable address yet. A project generated in Bolt runs first inside a sandboxed in-browser session, and while that session is enough for checking the work as it comes together, a launch listing benefits from a real deploy. Visitors who click through on launch day are not the only audience. People find Product Hunt listings weeks or months later through search, and a video pointing at a session that no longer resolves undermines a listing that is otherwise still doing its job.
Launch day itself adds a second reason to deploy before recording. Traffic on a listing can spike quickly once it starts climbing the rankings, and an in-browser preview session was never built to absorb that kind of attention even if it happened to still be running. A deployed app on a proper hosting target is the more reasonable target for whatever attention the launch actually brings, and the video should point at that same address from the start rather than needing to be swapped out mid launch.
