AI video continuation and generative extend
Add natural footage before or after a clip
AI Video Continuation adds matching footage before or after one short source clip, holding a moment longer, completing a reveal, adding a lead-in, or creating a natural ending. The extended clip comes back as an MP4 of up to 30 seconds, with the new footage generated from your clip's subject, camera, lighting, and atmosphere. AI Video Continuation works in Claude Code, Codex, Cursor, and other agents that run Skills.
Install AI Video Continuation and tell me when I need to sign in: https://beatra.ai/skills/ai-video-continuation/install.md
For agents:Install guideMarkdown version
Send it to WorkBuddy, Codex, Claude Code, or another agent to install. Sign in once the first time you use it.GitHub ↗Docs ↗
One source video, one naturally extended clip
Match the source boundary
Carry forward the subject, camera, lighting, and atmosphere at the join point so the new footage follows naturally.
One clear next beat
Extend toward one defined next event for a legible, reviewable continuation.
How it is made
Read the join point first
Capture subject, camera, lighting, and atmosphere at the source boundary, then define the single next beat.
Freeze subject, camera, lighting · Pick before or after · State the next beat
Generate the matching footage
Add one extension to the opening or ending and review the seam, motion direction, and continuity.
One source video, one extension · Review the seam · Deliver the extended clip
Stronger endings, longer holds, and action continuation
Stronger endings and longer holds
Hold a moment longer or land a more satisfying ending on an existing clip.
Action continuation and reveals
Carry an action forward or complete a reveal that the source cut short.
Product and landscape extensions
Extend a product reveal or landscape shot with matching footage for longer, richer clips.
Inputs and outputs
Input · Video — Short source clip
- Formats
- MP4, MOV
- Specifications
- 1 video per request; up to 100 MB each
Input · Text — The next moment to show
Output · Video — Clip with matching footage added before or after
- Formats
- MP4
- Specifications
- 1 video per request; 2–30 s; 480p, 720p, 1080p, 4K
| Input or output | Type | Formats | Specifications |
|---|---|---|---|
| Input | Video — Short source clip | MP4, MOV | 1 video per request; up to 100 MB each |
| Input | Text — The next moment to show | — | — |
| Output | Video — Clip with matching footage added before or after | MP4 | 1 video per request; 2–30 s; 480p, 720p, 1080p, 4K |
Specifications last checked · Sources: Model catalog · Documentation · Terms of Service
Questions
What do I need to start?
One short source video, whether to add footage before or after, and the next moment the new footage should reach.
What can I continue?
A stronger ending, a longer hold, the next action, a completed reveal, a lead-in, or any natural continuation of one source clip.
Will the new footage blend with my clip?
The new footage is generated from your clip and follows its motion direction, subject, camera, and lighting; the extended clip runs up to 30 seconds.
Related Skills
