One practical addition: record against seeded demo data on the checked-out tag, so the clip doesn't show someone's real workspace and can be regenerated identically if the script changes. Burned-in captions also help, since most people watch release videos muted. And if you want the clip to play inline, drag the MP4 into the release notes body rather than only attaching it as a release asset, because assets just show up as a download link.
DemoVexa is a desktop app that reads your code and turns it into finished product demo videos, with a script, voiceover and your real UI animated. No screen recording.
Good additions, all three. Seeded data on the checked-out tag is the one I'd push hardest, because it also keeps the clip matched to the code in that release instead of whatever is on main by the time someone records. Captions make sense for release videos in particular, since most people skim them in a feed with the sound off. And the inline MP4 tip is a good catch, an attached asset is easy to miss when it only shows up as a download link. Do you script the seed data per release, or keep one fixed demo workspace and only touch it when the UI changes?
I keep one fixed demo workspace and only touch it when the UI moves. The seed script runs against that workspace on the release tag, so the clip stays tied to what shipped without inventing a new dataset every time. When a feature needs fresh sample content, I add it to the seed rather than recording off a dirty local DB. The important bit is still "checked-out tag + seed", not whether the fixture is one file or many.
DemoVexa is a desktop app that reads your code and turns it into finished product demo videos, with a script, voiceover and your real UI animated. No screen recording.
That split makes sense. The workspace stays stable and the seed is the only thing that grows with features, so anyone on the team can rebuild the same clip later without needing your local database. One thing I'd wonder about is older releases: if you regenerate a clip for an old tag, does today's seed still run against it, or do you pin the seed version to the tag as well?
Yes — I pin the seed to the tag, same as the app. The seed lives in the same repo, so checking out v1.2 gives me the v1.2 seed and the v1.2 UI, and today's seed only comes in when we cut a new release. If a later seed change would break an old workspace, regenerating that old clip is "check out the tag, run the seed that shipped with it," not "run main's seed against an old checkout." The release video is a CI artifact of that tag, so we don't have to remember the ritual six months later.
DemoVexa is a desktop app that reads your code and turns it into finished product demo videos, with a script, voiceover and your real UI animated. No screen recording.
That's the cleanest version of it. Once the clip is a CI artifact of the tag, nobody has to remember how it was made, because the tag already carries the app, the seed and the recipe. It also answers the old release question completely: an old clip is just an old checkout plus the seed that shipped with it. Does the CI job block the release if the clip fails to render, or does it only warn and let the release go out?
We're a place where coders share, stay up-to-date and grow their careers.
One practical addition: record against seeded demo data on the checked-out tag, so the clip doesn't show someone's real workspace and can be regenerated identically if the script changes. Burned-in captions also help, since most people watch release videos muted. And if you want the clip to play inline, drag the MP4 into the release notes body rather than only attaching it as a release asset, because assets just show up as a download link.
Good additions, all three. Seeded data on the checked-out tag is the one I'd push hardest, because it also keeps the clip matched to the code in that release instead of whatever is on main by the time someone records. Captions make sense for release videos in particular, since most people skim them in a feed with the sound off. And the inline MP4 tip is a good catch, an attached asset is easy to miss when it only shows up as a download link. Do you script the seed data per release, or keep one fixed demo workspace and only touch it when the UI changes?
I keep one fixed demo workspace and only touch it when the UI moves. The seed script runs against that workspace on the release tag, so the clip stays tied to what shipped without inventing a new dataset every time. When a feature needs fresh sample content, I add it to the seed rather than recording off a dirty local DB. The important bit is still "checked-out tag + seed", not whether the fixture is one file or many.
That split makes sense. The workspace stays stable and the seed is the only thing that grows with features, so anyone on the team can rebuild the same clip later without needing your local database. One thing I'd wonder about is older releases: if you regenerate a clip for an old tag, does today's seed still run against it, or do you pin the seed version to the tag as well?
Yes — I pin the seed to the tag, same as the app. The seed lives in the same repo, so checking out v1.2 gives me the v1.2 seed and the v1.2 UI, and today's seed only comes in when we cut a new release. If a later seed change would break an old workspace, regenerating that old clip is "check out the tag, run the seed that shipped with it," not "run main's seed against an old checkout." The release video is a CI artifact of that tag, so we don't have to remember the ritual six months later.
That's the cleanest version of it. Once the clip is a CI artifact of the tag, nobody has to remember how it was made, because the tag already carries the app, the seed and the recipe. It also answers the old release question completely: an old clip is just an old checkout plus the seed that shipped with it. Does the CI job block the release if the clip fails to render, or does it only warn and let the release go out?