Step 1
Open Krea Video
Pick Runway Gen-4.5 or Gen-4 from the model selector in the Krea Video Tool.
Generate
Edit
Customize
"People running, risograph poster"
Generate
Edit
Customize
Generate AI videos with Runway Gen-4.5 on Krea — native text to video, image to video, consistent characters from one reference image, and realistic motion for ads, social clips, and film prototypes. Gen-4 is available in the same tool.
How It Works
From first prompt to production-ready video in four steps.

Step 1
Pick Runway Gen-4.5 or Gen-4 from the model selector in the Krea Video Tool.

2Step 2
Upload a photo or generate a reference image with Krea 1 — this frame locks your subject and composition.
A man painting in his garage, muted tones, high quality shot.
Step 3
Write a short prompt about how things should move. Focus on motion, not visual details.
Step 4
Compare takes, then enhance and upscale the best one — all inside Krea.
Runway’s latest frontier model with native text-to-video — live on Krea
Clip lengths on Gen-4.5 — Gen-4 generates 5 or 10 second clips
Smooth, film-style motion output
Characters stay consistent from a single reference across shots, lighting, and locations
Native text-to-video on Gen-4.5, plus start-frame animation across the family
Widescreen and vertical output for every placement
Typical Gen-4 generation time on Krea — Gen-4.5 takes around five minutes
Upload one reference image and your character keeps the same face, clothing, and colors in any lighting, location, or style. Build narrative content and short films without re-casting your subject every generation.
Generate with Gen-4.5Turn stills into motion — product shots, illustrations, or frames generated with Krea’s image models. Runway simulates real-world physics, so subjects move naturally instead of drifting or morphing.
Animate a Start FrameGen-4.5 generates video straight from a text prompt — no start frame required. Draft product videos, concept tests, and social clips in widescreen 16:9 or vertical 9:16 for TikTok, Reels, and YouTube.
Generate Video from TextPrompt Examples
Runway responds best to short prompts that describe movement instead of appearance — your reference image already shows what things look like. Every clip below is paired with the exact prompt that generated it.
A man painting in his garage, muted tones, high quality shot.
Try this promptPrompt
A fluid tracking shot following a lone figure walking through a rain-soaked dystopian street market at night, shot on Steadicam. The environment is be dense with neon lights (pinks, blues, purples), steam rising from grates, and hanging lanterns. The subject should be slightly underexposed, mysterious, with only rim lighting defining their silhouette.
Try this promptA fluid tracking shot following a lone figure walking through a rain-soaked dystopian street market at night, shot on Steadicam. The environment is be dense with neon lights (pinks, blues, purples), steam rising from grates, and hanging lanterns. The subject should be slightly underexposed, mysterious, with only rim lighting defining their silhouette.
Try this promptPrompt
Documentary, naturalistic shot. An old grizzled mechanic, somewhat overweight, wearing a plaid shirt, sits outside a garage at sunset, the warm light illuminating his face and beard. He's lost in contemplation. In the background we see sparks from welding. Slice of Americana documentary, New Mexico. Shallow depth of field, hard shadows.
Try this promptDocumentary, naturalistic shot. An old grizzled mechanic, somewhat overweight, wearing a plaid shirt, sits outside a garage at sunset, the warm light illuminating his face and beard. He's lost in contemplation. In the background we see sparks from welding. Slice of Americana documentary, New Mexico. Shallow depth of field, hard shadows.
Try this promptPrompt
Luxury Automobile - Cinematic Reveal - A sleek, black sports sedan emerges from darkness. Pull back slowly with a crane or jib to reveal the vehicle's sculptural curves. The car sits on a wet, reflective black surface in a minimal studio environment, lit with a single large soft box from above and slightly behind, creating elegant highlights along the body lines and hood. Position rim lights (hard sources with blue gels) on both sides to define the silhouette against the dark background. Grade with deep, crushed blacks and pristine highlights, emphasizing the premium craftsmanship. The shot should feel aspirational, powerful, and meticulously crafted.
Try this promptLuxury Automobile - Cinematic Reveal - A sleek, black sports sedan emerges from darkness. Pull back slowly with a crane or jib to reveal the vehicle's sculptural curves. The car sits on a wet, reflective black surface in a minimal studio environment, lit with a single large soft box from above and slightly behind, creating elegant highlights along the body lines and hood. Position rim lights (hard sources with blue gels) on both sides to define the silhouette against the dark background. Grade with deep, crushed blacks and pristine highlights, emphasizing the premium craftsmanship. The shot should feel aspirational, powerful, and meticulously crafted.
Try this promptPrompt
A dramatic time-lapse showing storm clouds rolling in over an open landscape or cityscape, captured over 2-4 hours compressed into 10-15 seconds. Highly sped-up footage. Set up a locked-off camera on a sturdy tripod in a location with an interesting foreground element (lone tree, building, or monument) and expansive sky. Shoot one frame every 3-5 seconds depending on cloud speed, maintaining consistent exposure despite changing light conditions. Frame with roughly two-thirds sky to emphasize the meteorological drama. Capture the progression from clear or partly cloudy to ominous storm clouds, including the dramatic shift in lighting as the sun is obscured.
Try this promptA dramatic time-lapse showing storm clouds rolling in over an open landscape or cityscape, captured over 2-4 hours compressed into 10-15 seconds. Highly sped-up footage. Set up a locked-off camera on a sturdy tripod in a location with an interesting foreground element (lone tree, building, or monument) and expansive sky. Shoot one frame every 3-5 seconds depending on cloud speed, maintaining consistent exposure despite changing light conditions. Frame with roughly two-thirds sky to emphasize the meteorological drama. Capture the progression from clear or partly cloudy to ominous storm clouds, including the dramatic shift in lighting as the sun is obscured.
Try this promptPrompt
3D render animation, a cute duckling character does a dance routine on a lilypad, in the style of modern 3D children's animated movies.
Try this prompt3D render animation, a cute duckling character does a dance routine on a lilypad, in the style of modern 3D children's animated movies.
Try this promptCopy the exact prompts behind these Runway clips and start generating your own AI videos.
Trusted by over 30 million creators worldwide.
Real reviews on TrustpilotSo many AI tools out there but I always come back to Krea. It's really the one AI platform that 'just works'.
FYN
Verified review
Krea's interfaces are still the best in the industry. Everything looks and feels so clean and easy to use! Imo the easiest to use gen AI website.
Lynn
Verified review
Ultra fast generations. Extremely simple to use. All the latest AI models.
Daniel
Verified review
KREA is the only AI subscription I have right now besides ChatGPT.
Sophia M.
Verified review
I'm a Krea Max user. It's crazy how fast they add models. I see a new AI model on twitter and the same day krea already offers it.
Wirot Ch.
Verified review
Still the most powerful AI creative suite out there.
Gravion
Verified review
Free
Get free daily credits to try basic features.
Basic
Access our most popular features
Pro
Advanced features and discounts on compute units
Max
Full access with higher discounts on compute units
For Teams and Enterprises
Workplace management, collaboration, and enterprise customizations
Business
Secure and collaborative workspace for growing teams
Enterprise
Enterprise-grade security with dedicated support and admin features
Runway Gen-4.5 is Runway’s latest frontier AI video model and the default Runway model on Krea. Its headline capability is native text-to-video — generate a clip from a prompt alone, with no start image — plus flexible clip lengths from 2 to 10 seconds, consistent characters, and realistic motion.
Gen-4 is the previous Runway generation, known for character consistency: upload one reference image and your character looks the same across different shots, lighting, and locations. Gen-4.5 builds on it with native text-to-video and more flexible clip lengths, while Gen-4 remains available on Krea for fast, image-driven generations.
Krea’s free plan includes daily compute units you can spend on any model, including Runway. For more generations, higher limits, and commercial licensing, upgrade to a paid plan.
On Krea, Runway Gen-4 generates 5 or 10 second clips, and Gen-4.5 supports any length from 2 to 10 seconds. You can choose widescreen 16:9 or vertical 9:16 formats.
Gen-4 clips typically generate in about a minute on Krea. Gen-4.5, the higher-quality frontier model, takes around five minutes per clip.
Keep prompts simple and motion-first. Start with basic movement, then add details one at a time. Describe how things move rather than what they look like — your start image already covers appearance. Avoid conversational language and negative prompts.
Use any image as your starting frame. Generate one with Krea 1 or another model in the Krea Image Tool, then switch to Runway in the Video Tool — or upload your own photos and illustrations.
Yes. Runway Gen-3, the previous-generation cinematic model, is still available on Krea in the Runway Gen-3 video tool. Gen-4 and Gen-4.5 supersede it for character consistency and prompt adherence.
Yes. On Krea paid plans, generated videos include commercial licensing for campaigns, products, client work, and other production projects.
Runway keeps characters consistent within one clip, but it can’t maintain the same character across multiple separate clips — each generation is one scene. Avoid prompts that ask for multiple scene changes or style shifts in a single clip.
Latest news, tutorials, and updates from the Krea team

How we make the default Kubernetes scheduler pack GPU workloads into the cluster first and spill to overflow capacity only when it's full, using dynamic taints, the Descheduler, and Kueue.

How Taskin Goec rebuilds fashion's most expensive steps with Krea.

How Krea shares GPUs between research and production with a custom solution that lets inference scale anywhere when training claims every GPU.
Generate AI videos with Runway Gen-4.5 on Krea. Create clips straight from a text prompt with native text-to-video, keep characters consistent from a single reference image, and ship widescreen or vertical formats — with Gen-4 also available in the same tool.
Try Runway on KreaExplore other Runway models available on Krea
Discover more AI models for images and videos on Krea
Or try every model side by side in Krea's free AI video generator and AI image generator.
The Gen video model family is developed by Runway. For official model announcements and research, visit the provider’s site.
Visit the official Runway sitePage last updated