Turn Screenshots Into a Video Without a Timeline Editor
Select the images, right-click, and get one captioned, animated .mp4 back — the screenshots were already sitting in the folder.
Summary
Putting together a short intro or demo clip usually means opening a separate video editor, importing images one at a time, adding captions by hand, and figuring out transitions — for images that were already sitting in the folder.
The Video Composer plugin skips the editor. Select the images in Filer, right-click, and either write a caption per image or let your configured model draft them — the plugin composes them into one Ken-Burns-animated .mp4, with a matching .srt sitting right next to it.
Follow along
The screen below is from an actual run of the plugin — the same one used to compose the intro video on this site's own home page.
Needs: the Video Composer plugin installed (it bundles its own ffmpeg — no separate install). AI-drafted captions additionally need a configured Chat AI, local or remote.
Select the images, right-click
Multi-select the images in the order you want them to play, right-click, and open Actions → Create Video...
Write the captions, or have AI draft them
The dialog asks for a caption per image, how many seconds each scene stays on screen, and where to save the result. Filling every caption by hand is one option — the Draft Captions (AI) button is the other: give it a one-line description of what the video is for, and your configured model writes a caption per image from that description (it works from the description, not by looking at each image's content — write one that gives it enough to go on).
Video intro/purpose: "A 60-second tour of Filer's Explorer, AI Actions, and folder agents"
Send it, get back an .mp4 + .srt
Each image becomes an animated scene with its caption burned in, back to back into one video, with a sidecar .srt for anyone who wants to read along instead.
One frame from the actual finished video this workflow produced — not a mockup.
Other ways to reach the same result
From chat — Draft Captions and Create Video are ordinary tools the plugin exposes, not a UI-only feature, so asking a folder's agent in chat to "combine these three screenshots into a video" is a real path in if the model decides to call the tool — just not the one this walkthrough verified end to end.
On a schedule or a file event — the same tool call works from a scheduled or file-triggered agent, not only a one-time right-click: a folder that watches for new screenshots could compose an updated highlight reel on its own each time one lands.
This isn't a mockup — the frame above comes from an actual 60-second video the plugin composed end to end, the same one now used on this site's own home page: real screenshots, a real Ken-Burns render per scene, and real captions burned in by the plugin itself.
Where else this applies
Any folder that piles up images faster than anyone turns them into something presentable fits the same shape:
- Product or feature walkthroughs — the screenshots for a release note or a support answer, turned into a short clip instead of a wall of images.
- Before/after documentation — a handful of progress screenshots from a project folder, composed into a short recap without opening an editor.
- A lighter-weight alternative — when a full captioned video is more than the moment needs, a plain looping GIF from the same kind of image selection is the smaller, no-caption option.
What it needs
- The Video Composer plugin installed — it bundles its own ffmpeg, no separate service or install
- 2 or more images (.png, .jpg, .jpeg, .webp, .bmp) selected in the order they should play
- An inference endpoint — only if using Draft Captions (AI) instead of writing captions by hand