AI image and video generation and editing for motion, storytelling, and creative workflows.
ImageVideoAudio
AI image and video generation and editing for motion, storytelling, and creative workflows. Open app.runwayml.com, pick a generative mode (Text to Video, Image to Video, and others), draft a prompt focused on motion (up to 1000 characters for Gen-4 video), and generate in 5- or 10-second clips. Gen-4.5 is the recommended model for quality; Gen-4 uses 12 credits per second, Gen-4 Turbo uses 5, and Gen-4.5 uses 12 per official help.
Open Text to Video with Gen-4.5. Describe what appears in frame and how it moves using clear natural language. Start simple, then add detail.
What to provide
Details that improve the result
Example prompt
Slow tracking shot through a sunlit kitchen. Steam rises from a mug on the counter. Morning light through the window. Camera moves forward gently. No dialogue, no logos, no text.
If the first output is not good
Common mistakes
Upload the image as the first frame. Write a prompt almost entirely about motion: subject action, camera move, environmental reaction.
What to provide
Details that improve the result
Example prompt
Image: portrait of a woman in a raincoat. Prompt: “The camera slowly pushes in as the subject turns toward the streetlight. Rain streaks move down the frame. Keep the same color grading.”
If the first output is not good
Common mistakes
Open Runway’s image generation tool, describe one scene with concrete visual details, generate, then use the still as an Image to Video input if needed.
What to provide
Details that improve the result
Example prompt
Flat illustration of a home office at night. Warm desk lamp, closed laptop, notebook and mug, deep blue background, minimal shapes, no text, no watermark. 16:9.
If the first output is not good
Common mistakes
Open Video to Video (Edit Studio / Aleph). Upload clip, describe the desired change as a delta, generate, and iterate on the closest pass.
What to provide
Details that improve the result
Example prompt
Upload a walking-down-the-street clip. Prompt: “Convert to animated watercolor style. Preserve walking pace and camera angle. No added characters.”
If the first output is not good
Common mistakes
Use the Animate Frames app to interpolate between uploaded or generated keyframes. Describe how the subject and camera move from frame A to frame B.
What to provide
Details that improve the result
Example prompt
Start frame: product on white table. End frame: same product, camera closer, slight rotation. Prompt: “Smooth dolly in with gentle product rotation between keyframes.”
If the first output is not good
Common mistakes
Use the Runway Dev API for programmatic generation. API credits bill separately from app.runwayml.com subscriptions per official FAQs.
What to provide
Details that improve the result
Example prompt
API Image to Video: upload portrait still, prompt “Slow zoom in as the subject smiles subtly”, duration 5s, ratio 16:9, model Gen-4 Turbo for lower credit cost.
If the first output is not good
Common mistakes
Runway prompts should describe motion clearly. For Image to Video, the image defines the look—your text describes what moves. For Text to Video, include both visuals and motion. Start simple, iterate one variable at a time.
Mode: Image to Video. Image: uploaded still of cyclist on a bridge. Prompt: “The camera tracks alongside as the subject pedals. City lights blur in the background. Locked exposure.” Duration: 5 seconds.
The input image already conveys subject, lighting, and composition. Describe subject action, camera move, and scene motion instead.
Example
“Handheld camera: the subject stands still as the crowd moves around them. Natural camera shake.”
Text to Video needs both what appears and how it moves. Natural language sentences reduce ambiguity.
Example
“Wide shot of a fox in snow. The fox lifts its head and exhales visible breath. Camera holds steady, soft morning light.”
Use “the subject”, “she”, or “the object” so the model focuses on motion rather than reinterpreting image details.
Example
“The subject turns slowly toward camera” instead of re-listing clothing and hair from the image.
Runway recommends a foundational prompt first. Add camera complexity, environmental motion, or timing after reviewing the first clip.
Example
Pass 1: “Slow push in.” Pass 2: “Slow push in as steam rises from the mug.” Pass 3: add “handheld shake” only if needed.
Describe what should happen and what should stay still. Positive instructions work better than long negative prompt lists.
Example
“Keep the background static and color grading unchanged” instead of a long list of exclusions.
Pick the right generative mode
Text to Video and Image to Video use Gen-4.5 for highest quality. Gen-4 and Gen-4 Turbo trade cost and speed. Video to Video edits existing clips; Animate Frames interpolates keyframes.
Treat the image as frame one
For Gen-4 Image to Video, the upload defines composition, lighting, and style. Prompt almost entirely for motion, camera, and temporal change.
Describe camera motion explicitly
Locked, handheld, dolly, pan, tracking, whip pan, zoom, and focus shifts are valid prompt components. One primary camera move per test is easier to debug.
Separate subject and scene motion
Subject motion covers characters and objects. Scene motion covers environment reactions—rain, crowd, foliage, particles.
Match duration to motion complexity
Gen-4 supports 5- and 10-second outputs at 24fps. Complex motion may read cleaner in 5 seconds before extending.
Choose aspect ratio early
Official outputs include 16:9, 9:16, 1:1, 4:3, 3:4, and 21:9 resolutions. Set ratio before generating to avoid reframing later.
Use Gen-4 Turbo for iteration
Gen-4 Turbo costs 5 credits per second versus 12 for Gen-4 and Gen-4.5—useful for prompt iteration, then switch to Gen-4.5 or Gen-4 for final quality.
Generate reference stills first
Image models share the credit wallet with video on Standard and above. A strong still reduces visual surprises when you animate.
Avoid negative prompt sprawl
Runway recommends positive phrasing. State desired motion and what must stay unchanged instead of long exclusion lists.
Keep text prompts under the limit
Gen-4 text prompts allow up to 1000 characters. Clarity beats length—prioritize the most important motion components.
Plan credits before batches
Credits spend per second based on model and duration (e.g., Gen-4 at 12 credits/s vs Gen-4 Turbo at 5 credits/s per official help).
Separate app and API billing
Runway Dev API credits bill separately from app.runwayml.com subscriptions. Check API FAQs before integrating.
Iterate from the closest clip
Regenerate from the best prior output with one changed instruction instead of restarting from a unrelated prompt.
Use natural language
Full sentences give context for how elements appear and move. Keyword stacks alone increase ambiguity in Text to Video.
Check mobile compatibility
Gen-4 is available on web and the iOS app per official help. Confirm device support before planning a mobile workflow.
Pricing, features, limits and latest updates
AI image and video generation and editing for motion, storytelling, and creative workflows.
Free / Paid · Free
Free
Free
Standard
$15/ Monthly
Standard
$144/ Yearly
Pro
$35/ Monthly
Pro
$336/ Yearly
Max
$95/ Monthly
Max
$912/ Yearly
Enterprise
Custom
Image models share the same credit wallet as video on Standard and above.
Generative video models such as Gen-4.5 spend monthly credits on paid plans.
Runway Dev API credits are billed separately from app.runwayml.com subscriptions.
Official pricing lists Lyria song generation billed in credits.
Pro and higher include custom voices for lip sync and text to speech.
Share a prompt that worked for you. Username and email are shown with your submission. External links are not allowed.
Example prompt
Prompt
A slow tracking shot through a sunlit kitchen. Steam rises from a mug on the counter. Natural morning light. Gentle camera push-in. No dialogue, no logos.
Short explanation
Runway prompting guidance focuses on motion and camera behavior; describe one clear shot per generation.
Example prompt
Prompt
[upload still image] Motion: subtle leaf movement in the foreground, soft wind, camera holds steady. Keep the subject unchanged. No scene cuts.
Variation
Same image. Add only light flicker from a nearby screen; everything else static.
Short explanation
For image-to-video, prompts should describe movement while preserving the reference composition.
Required fields are marked. Variation, result, and explanation are optional.
Image generation with accurate in-image text, typography, and layout.
ImageFree / Paid
Generative AI for images, video, audio, and design built for commercially safe creative work.
ImageFree / Paid
explore