01
Choose a clear reference
Short clips with deliberate camera work and visible action produce more focused prompts than long montages.
Video to prompt generator
Turn a reference clip into an AI video prompt.
A useful video prompt describes more than the subject. AdScanVideo breaks reference footage into shots and records the camera, movement, light, timing, and visible action needed to recreate or adapt the idea.
No signup · One free analysis daily during early access · Files removed after processing
Shot by shotstructured prompts
Timedchronological beats
Model-readycopy and adapt
$3.99early-access paid analysis
The prompt should preserve the directing decisions
“A person holding a product” is a subject description, not a production prompt. A stronger prompt explains where the camera is, how it moves, what the subject does over time, how the scene is lit, what should remain consistent, and how the shot ends.
AdScanVideo extracts those decisions from a reference clip. You can then adapt the result for Sora, Veo, Kling, Runway, Seedance, or another video model. Every generator interprets prompts differently, so use the result as a precise starting draft and adjust it for the model you choose.
02
Generate a shot breakdown
The analyzer describes the sequence, framing, composition, light, motion, and dialogue that can be inferred from sampled frames.
03
Adapt instead of copying
Change the subject, setting, product, wardrobe, branding, and story purpose while preserving the useful camera language.
Example prompt structure
This example shows the format and level of detail. Actual results depend on the video, selected mode, captions, audio quality, and visible evidence.
SHOT 03 · 00:06–00:09 Medium close-up of a runner stopping beneath a streetlight at blue hour. Handheld camera moves forward slowly at chest height. Cool ambient light contrasts with a warm pool of light on the face. The subject catches their breath, looks toward camera, and raises a water bottle into frame. Shallow depth of field, natural skin texture, restrained documentary color grade. End on the bottle label held steady for half a second.
See the difference on a real eight-second clip
Our published shot-breakdown sample describes an opening 00:00–00:01 moment: an orange cartoon cat and a duck in a grassy field under a blue sky, with a slight camera zoom. That is an observation from a processed video, not a generated video prompt.
To turn a similar camera idea into an original prompt, keep the shot structure and change the characters and setting. For example:
Vertical medium shot, 1 second. Two original paper-cutout characters stand in a sunlit rooftop garden. Begin with both figures in frame and move the camera forward in a subtle, continuous zoom. Bright blue sky, soft cloud shapes, clean readable silhouettes. End before the next scene change.
This rewritten prompt is a human-written illustration, not verbatim analyzer output. Check the real report for the timestamped visual evidence, then edit your own prompt for your chosen video model.
Prompt details to keep
- Subject and continuous action
- Shot size, angle, and composition
- Camera movement and apparent speed
- Lighting direction, color, and mood
- Duration, dialogue, and transition
Details to rewrite
- Brand and product identity
- Protected characters or distinctive artwork
- Claims and on-screen copy
- Location and wardrobe for your concept
- Platform-specific aspect ratio and duration
The analysis describes visible creative choices. It does not reproduce an original video or guarantee that a generation model will match it exactly.
Questions before you analyze
Can a video be converted into a text prompt?
Yes. The analyzer samples the video and creates timestamped descriptions of subject, action, framing, camera movement, lighting, and style that can be adapted into prompts.
Which AI video models can use the prompts?
The output is plain structured text, so it can be adapted for Sora, Veo, Kling, Runway, Seedance, and other prompt-based video tools.
Does the prompt include dialogue?
It includes dialogue when captions or a usable audio transcript are available. Visual-only footage receives visual prompts without invented speech.
Can I recreate any copyrighted video?
Use references you own or are authorized to study. Transform the idea into original work and respect copyright, trademarks, publicity rights, and platform rules.
Related workflows
Turn the next video into something you can use.
Upload a file or paste a public link. Choose the focused mode and get a structured report with one free analysis daily during early access.
Start free analysis