The real opening image
Download the START frame captured at the beginning of the shot, rather than a midpoint image that already contains later action.
Runway workflow selected
Video to Prompt for Runway
Upload a reference clip and get one Runway-ready motion prompt plus the real opening frame for every detected shot.
- Runway selected before upload
- Real first frame by shot
- Short motion-focused prompts
Upload and analyze
Choose a model, then upload your video
Get prompts by shot, real start and end frames, and model-ready instructions in one analysis.
2. Upload your source video
Visual detailCamera motionSound notes
Choose a short video to begin.
Report a problemVideo previewWaiting for upload
Your source stays visible
Compare every generated shot description with the original video before you copy the prompt.
Keyframes
Editable prompts
Audio review
Direct answer
Use the image for appearance and the prompt for motion
For Runway image-to-video, upload the actual first frame of one shot. Let that image carry the subject, composition, lighting, color, and style. Use the text prompt to say what the subject, environment, and camera do next, including direction, speed, and timing when those details are visible.
Generate one shot at a time. A clip with three cuts needs three starting images and three motion prompts, followed by a simple edit that restores the source order and timing.
Less text, clearer job
What the Runway Prompt Pack Gives You
This page preselects Runway before the upload. The same video analysis remains editable, but the final prompt removes details already visible in the starting image and focuses on movement.
Subject motion
The prompt states the physical action: turns, walks, slides, grows, falls, looks left, or remains still.
Scene and camera motion
Environmental movement and camera behavior are added only when the source shows them. A fixed shot is written positively as a locked or still camera.
Timing and edit notes
The export keeps source timecodes, generation-length guidance, shot order, exact overlay text, and sound notes outside the prompt box.
A repeatable correction loop
How to Use a Video Prompt in Runway
The first attempt is a test of one motion instruction. When it fails, change the smallest field that explains the miss rather than making the whole prompt longer.
- 01
Check where the shot begins
Play the source at the listed start time. Confirm that the downloaded START frame matches the composition before any key action has already happened.
- 02
Upload that frame to Runway
The image becomes the visual starting point. Use a clean source frame because blur, broken hands, or compression artifacts can carry into the generated motion.
- 03
Paste only the motion prompt
Copy the selected shot’s prompt. Keep the central subject action, then include camera or environmental motion only when it materially changes the shot.
- 04
Compare the generated motion
Check direction, speed, camera behavior, and action order against the reference. If one element is wrong, edit that element and run another attempt.
- 05
Trim and finish outside the model
Keep the source-length portion, add exact on-screen text and sound in an editor, and place the shot before or after the neighboring generated clips.
Verified fixture pattern
Turn a Long Description Into a Runway Motion Prompt
In the project's checked cup fixture, the opening image already shows a yellow cup, the dark blue background, the centered composition, and the word “MUSE.” Repeating those facts in the Runway text box adds little. The change that must happen is the cup becoming larger while the view stays fixed.
Official guidance checked September 15, 2026
What Runway Says the Prompt Must Control
Runway's current image-to-video guide says the input image defines composition, subject matter, lighting, and style. The text should focus on what happens: subject action, environmental motion, camera motion, movement style, timing, direction, and speed. It recommends starting with the most important motion and adding detail only when needed.
| Source evidence | Put it where? | Example |
|---|---|---|
| Subject appearance and composition | START image | The actor, product, framing, colors, lighting, and background are already visible. |
| Physical subject action | Prompt | The subject turns slowly and reaches toward the cup. |
| Camera movement | Prompt when visible | The camera tracks left with the subject, or the locked camera remains still. |
| Exact title and soundtrack | Edit notes | Rebuild readable text and restore dialogue, music, or effects after generation. |
Capability sources: Runway Image to Video Prompting Guide and Creating with Gen-4.5. The exported pack currently labels the 5/10-second Gen-4 and Gen-4 Turbo workflow; always confirm the model selected in your own Runway workspace.
Fit and limits
When a Runway Motion Prompt Is the Right Output
Use this page when you already have a strong source frame and need to reproduce a clear movement: a product rotates, a person turns, fabric moves in wind, a camera pushes in, or the scene stays locked while one object changes.
It is a weaker fit when a shot must land on a precise final composition, because this workflow supplies only the opening frame. For a controlled start-to-end transition, compare the Veo first-and-last-frame workflow.
Neither workflow reveals the original prompt or guarantees an exact copy. Frame to Prompt shortens the work before your first generation and makes the next correction concrete: change the frame when appearance is wrong, or change the prompt when motion is wrong.
Specific questions
Video to Prompt for Runway FAQ
How do I turn a video into a Runway prompt?
Choose Runway, upload the reference clip, and review each detected shot. Download that shot’s START image, then paste its short motion prompt into Runway image-to-video. Generate and trim one shot at a time.
Why is the Runway prompt shorter than the Veo prompt?
In Runway image-to-video, the starting image already provides subject appearance, composition, color, lighting, and style. Repeating all of those details in text can distract from the main job of the prompt: describing motion and time.
Which frame should I upload to Runway?
Use the START frame captured at the beginning of the shot. A middle keyframe may already show the action in progress and can change the opening state, so Frame to Prompt labels the three frame roles separately.
What should a Runway image-to-video prompt include?
Start with the subject’s physical action, then add environmental motion, camera movement, direction, speed, or timing only when the source makes them clear. Use direct, positive language and keep one shot in one request.
Can I paste the entire exported pack into Runway?
Do not paste the production notes, shot order, frame instructions, and uncertainty section into the model. Paste only the text between the prompt markers for the selected shot.
Can this recreate exact text and audio in Runway?
Generated typography and sound may not match the source. The pack records visible words and audio cues for your edit, but the Runway motion prompt stays focused on the visual action. Rebuild exact titles and sound in an editor when needed.