Everything you need — models, prompting, pricing, camera moves, aspect ratios, workflows, and a newbie start guide — all in one place.
Runway is an AI research company building General World Models — AI systems that can understand, perceive, generate, and act in the world. Their consumer platform is the leading tool for AI video generation, editing, and motion capture used by filmmakers, studios, advertisers, game developers, and solo creators worldwide.
Text-to-video and image-to-video using Gen-4.5, the world's top-rated video generation model in 2026.
Edit existing video with text prompts — swap backgrounds, add objects, change lighting, generate new angles.
Transfer facial expressions, body movement, and performance from any video to a digital character. No suit needed.
Add speech, lip-sync, and audio to still images and videos using AI-generated voice tools.
GWM-1 powers real-time simulation, explorable environments (Worlds), and conversational video avatars (Characters).
Build automated, multi-step AI pipelines for character generation, storyboarding, and creative production.
Never used Runway before? Start here. This is the exact path from zero to your first real video generation in under 30 minutes.
Go to app.runwayml.com and sign up with Google or email. No credit card required for the free tier. You'll land on the main dashboard. You have 125 free credits to explore — treat them as a test drive, not a workspace.
The main dashboard shows your Projects, a Generate button for new creations, your credit balance in the top right, and quick access to recent outputs. The left sidebar has Gen-4.5, Aleph, Act-Two, and Workflows. Explore but don't generate yet — save your free credits for something intentional.
Click Generate → Video → Gen-4.5. You'll see a text prompt box and settings for aspect ratio (16:9, 9:16, 1:1) and duration (5 or 10 seconds). For your very first generation, try a simple, specific scene. Don't start with a vague abstract concept.
"A slow dolly shot through a misty forest at golden hour. Sunbeams filter through tall pine trees. Cinematic color grading. Shallow depth of field."
This is where Runway truly shines. Upload any image (a photo, an AI image, concept art) and write a prompt describing only the motion. The image already defines the visual — you're only directing what moves. This mode gives far more control than text-to-video for most creative projects.
Generation in Runway is a conversation, not a vending machine. Your first result won't be perfect — and that's normal. Take what you get, identify what's wrong (wrong motion? wrong framing? wrong vibe?), adjust one thing in your prompt, and regenerate. Mastery is built through iterations, not single attempts.
Go to academy.runwayml.com — it's completely free. Start with the Prompting Guide, then take the beginner course "AI for Advertising" (10 modules). The Academy has 31 courses and tutorials covering VFX, game dev, advertising, animation, and more — all taught with Runway's actual tools.
If you're serious about using Runway, Standard at $12/mo unlocks everything — Gen-4.5, Aleph, Act-Two, watermark removal, all third-party models. Your 625 credits/month = ~25 seconds of Gen-4.5. Budget your credits strategically: iterate with Gen-4 Turbo (5 cr/sec) and finalize with Gen-4.5 (25 cr/sec).
Runway offers a multi-model ecosystem. Here's every model you need to understand — what it does, when to use it, and what it costs in credits.
The world's top-rated video model in 2026. Gen-4.5 delivers unprecedented visual fidelity, cinematic realism, and precise camera control. It supports both Text to Video and Image to Video workflows and is what Runway is benchmarked on in industry comparisons.
Best for: Cinematic scenes, narrative content, advertising, any work where visual quality is the priority.
The reliable workhorse. Gen-4 is excellent for image-to-video workflows and uses the References system for character/environment consistency across scenes. It's the foundation of most production workflows when cost efficiency matters.
The iteration model. Gen-4 Turbo is visibly lower quality than Gen-4.5, but it's 5x cheaper per second. Smart creators use Turbo for rapid iteration and concept testing, then switch to Gen-4.5 to finalize their best shots. On the Unlimited plan, Turbo is available in Explore Mode with no credit cost.
The video editor that speaks English. Aleph lets you transform existing video footage with text prompts. Add or remove scene elements (backgrounds, props, characters), change environments and lighting, generate new camera angles, create green screen assets, and extend shots — all without re-generating from scratch. Aleph 2.0 now supports keyframe images for even more precise control.
Key capabilities: add scene elements, change environments, generate new angles, generate next shots, transform videos, create clean plates/green screens.
Democratized motion capture. Act-Two transfers facial expressions, body gestures, head movements, and even voice from any video to a digital character — using nothing more than your smartphone camera. Next-generation mocap that would have required a studio and a $50K suit is now a button click. You can also change character voices directly from within the interface.
Runway's most ambitious model. GWM-1 is a state-of-the-art General World Model that simulates reality in real time. It powers three groundbreaking products: GWM Worlds (open-ended interactive explorable environments), GWM Avatars / Characters (real-time conversational video agents with natural awareness), and GWM Robotics (physical interaction simulation for robotic applications).
Runway is more than a video generator. Here's the full feature set broken down by what it does and how to access it.
Generate video entirely from a written prompt. Supports Gen-4.5, Gen-4, and Gen-4 Turbo. Best results come from detailed motion descriptions and camera direction.
Upload an image as the first frame, then write a motion prompt to animate it. Your most powerful mode for control and visual consistency.
Upload reference images to lock in specific characters, locations, or objects across multiple generations. Essential for narrative and multi-shot work.
Edit real or generated video with text prompts. Change lighting, style, backgrounds, and scene elements without regenerating the whole clip.
Generate different camera angles from a single source shot. Create wider or tighter framings from existing footage — useful for coverage you didn't capture.
Turn any video into a green screen asset or clean plate. No rotoscoping required. Eliminates hours of manual masking work.
Film yourself or anyone performing, then transfer that performance — face, body, hands — to a digital character of any style.
Generate speech and add lip-sync to still portraits and videos. Custom voice creation available on Pro tier and above.
Upscale generated videos to higher resolution for final delivery. Available on all paid plans with watermark removal included.
A conversational generation interface — generate anything using Gen-4 images, videos, and references from within a single chat UI. Available to all users.
Full-featured timeline video editor built into Runway. Combine clips, adjust timing, and edit your AI-generated content without leaving the platform. Aleph 2.0 works inside Edit Studio.
Build automated multi-step AI pipelines. Featured workflows include Character Creator, Storyboard Creator, and Story Panels. Advanced users can build fully custom pipelines.
Programmatic access to Gen-4.5, Gen-4 Turbo, Aleph 2.0, and image models. Used by consumer tech companies to serve millions of generations. Enterprise-grade reliability.
Real-time video agent API for building fully custom conversational characters. Any appearance, any style, full voice and personality control. Zero fine-tuning from a single image.
Train custom models on your own visual style, brand, or characters. Used by studios like Lionsgate for bespoke AI production pipelines.
Runway uses a tiered subscription + credit system. Credits are the currency of generation — and the cost per second varies significantly by model. Know the math before you commit to a plan.
| Model | Cost | 625 cr = how much video | Best Use |
|---|---|---|---|
| Gen-4.5 | 25 cr/sec | ~25 seconds | Final output, hero shots |
| Gen-4 | 12 cr/sec | ~52 seconds | Production quality, lower cost |
| Gen-4 Turbo | 5 cr/sec | ~125 seconds | Rapid iteration, draft review |
| Gen-4 Image Turbo | 2 cr/image | ~312 images | Image generation, references |
Prompting is a new skill built on communication you already have. The key difference: generative models interpret your words literally and have no shared context. Be specific. Be concrete. Direct the camera.
[Camera movement] shot of [subject/object] [action] in [environment]. [Supporting details].
"A low-angle tracking shot of a woman in a black trench coat walking through a rain-soaked neon alley at midnight. The camera moves forward smoothly at walking speed. Steam rises from vents, reflections ripple across wet pavement, distant traffic glows in the background. Cinematic color grading, subtle handheld energy."
The camera [motion description] as the subject [action]. [Additional motion details].
"The camera executes an aggressive, sweeping horizontal arc around the subject, followed by an extremely rapid crash zoom that concludes with a sharp focus on the subject's eyes."
[00:01] Subject stands still. [00:03] Camera pushes in slowly. [00:05] Subject turns to face camera, smiling.
"A [shot type] of [subject] in [setting]. The camera [movement type] as [subject motion]. [Lighting description]. [Style/mood]. [Additional environmental details]."
"A slow circular dolly shot around [product] on [surface]. The camera orbits at [height]. [Lighting setup, e.g. soft studio lighting with warm rim light]. [Material behavior, e.g. glass surface reflects ambient light]. Clean and minimal background. Commercial photography aesthetic."
"Handheld vertical shot. [Subject] doing [action] in [location]. Natural lighting. Authentic, unpolished energy. Documentary feel. Shot on phone. [Any specific motion detail]."
"The camera slowly pushes in toward the subject's face. The scene comes to life with [primary motion, e.g. hair moving in breeze]. In the background, [secondary motion, e.g. leaves gently moving]. Minimal camera movement at the end to allow the scene to breathe."
Runway's models respond to cinematography vocabulary. The more precisely you direct the camera, the more cinematic and controlled your output. These are the terms that actually work.
Combining terms is encouraged — especially for cinematic sequences. Mix camera angles, motion style, and composition freely.
"Low angle, handheld tracking shot following the detective as he runs through a neon-lit rain-soaked alley. The camera keeps pace with him, with natural shake. As he rounds the corner, the camera whip-pans to a slow dolly push toward the door at the end of the corridor."
Runway supports multiple aspect ratios depending on the mode and model you're using. Choose your ratio based on where the content will live.
Runway Workflows let you build multi-step AI pipelines — automating complex creative processes and creating consistent, repeatable results at scale.
Generate consistent characters across multiple shots and styles. Uses References + Workflows to maintain visual coherence.
Build full visual storyboards from a concept. Generate sequential story panels with consistent world and character.
Build an entire world from a single reference image. Create consistent sequences across multiple shots using one source.
Build fully custom automated workflows for your production needs. Advanced users can chain models, set conditions, and automate entire production stages.
Draft all shots using Gen-4 Turbo (5 cr/sec). Test camera moves, timing, and compositions without burning your Gen-4.5 budget. Get the shot roughly right first.
Once you've locked your shot and prompt, generate the final version in Gen-4.5. Only use your flagship credits on shots you know are working.
Use Aleph 2.0 to refine the generated footage — adjust lighting, swap backgrounds, clean up problem areas, or generate coverage shots from existing clips.
For shots longer than 10 seconds: scrub to the last frame of your clip, select "Use current frame," and use it as the input image for the next generation. Combine in Edit Studio.
Capture facial performance, body movement, or dialogue on any camera. Use Act-Two to transfer the performance to your digital characters for the final sequence.
Runway is used across industries by professionals who need cinematic AI video in their workflow. Here's how each sector uses it.
Pre-visualization, VFX shot development, short film production, concept pitching. Used by studios in partnership with Lionsgate and others.
Product shots, social ads, campaign content, brand videos. Turn a single product photo into a full social media ad in minutes. AI for Advertising course available free.
Cinematic cutscenes, concept art animation, character movement, world-building assets. AI for Games course available free in Academy.
Product visualization, concept exploration, UI/UX motion references, design presentation videos.
Animated architectural walkthroughs from static renders. KPF uses Runway to animate projects in-house, replacing expensive external rendering pipelines.
Lookbook content, garment motion studies, editorial content, runway (literal) simulations. Generate diverse model representations with consistent styling.
UCLA Film, TV and Digital Media uses Runway to empower students to experiment with AI in filmmaking. Academic pricing available.
Client presentation content, pitch decks, rapid ideation and concepting, multi-channel campaign asset creation at scale.
These are the non-obvious things that separate creators who get great results from creators who struggle. Most people don't discover these until months in.
Motion blur, dust clouds, mid-action poses, and directional lines all act as motion cues. If your image shows a car with motion blur and you prompt for it to be stationary, you'll fight the model for multiple iterations. Remove motion cues from the source image first using text/image-to-image tools.
If you're getting unwanted jump cuts, increase duration. Cuts often happen when the model runs out of "space" to execute your sequence within the chosen duration. A complex sequence needs 10 seconds; a single camera move can work in 5.
Runway clips max out at 10 seconds. Extract the exact last frame (use the scrubber), drop it as an image input, and continue. The model will match the visual style and motion direction. You can build sequences of any length this way.
The model speaks cinematography. "Lens height, movement, motion beats, lighting geometry" produce better results than "emotional, atmospheric, evocative." Replace feelings with shots. Replace moods with lighting setups.
Without References, your character's face will change from shot to shot. Upload a clear, well-lit reference image of your character and use the @ recall feature in Gen-4 to maintain consistency across every generation in your project.
Most beginners only generate new clips. But Aleph lets you fix existing clips without starting over. Bad background? Fix it. Lighting off? Fix it. Need a wider angle? Aleph generates it. Treat it as a cost multiplier — one good generation can be edited 5 different ways.
Multi-paragraph prompts with every detail specified often produce worse results than shorter, clearer ones. The model struggles to honor every simultaneous constraint. Pick the 3-4 most important details and let the model fill in the rest with creative freedom.
1 second of Gen-4.5 = 5 seconds of Gen-4 Turbo in credits. Develop your shots entirely in Turbo. When you have the prompt and composition exactly right, generate that one shot in Gen-4.5. Your credit budget goes 5x further this way.
Text rendering inside video is still inconsistent (signs, labels often have errors). Hand/fine motor detail can produce artifacts. Very specific product details need image-to-video with real product photography rather than text-to-video. The API currently provides access to Gen-4 Turbo and Gen-4 Image, not Gen-4.5.
Academy.runwayml.com has 31 free courses and tutorials. Start with "AI for Advertising" (beginner, 10 modules) and "AI for Visual Effects" (intermediate, 6 modules). Watch the behind-the-scenes tutorials — they show real production workflows from real creators.
Academy.runwayml.com hosts 31 courses and tutorials — all free. Here are the most valuable ones and who they're for.
10 modules. Product shots, social ads, campaign content. Start here if your goal is commercial content creation.
6 modules. Integrate AI into VFX pipelines, pre-visualization, and production-ready shots.
5 modules. Asset creation, cinematics, and game production pipelines with Runway tools.
5 modules. Add scene elements, change environments, generate new angles, transform videos with natural language.
2 modules. Capture facial expressions, body gestures, and change character voices for dynamic animation.
3 modules. Build automated AI pipelines for character generation, storyboarding, and multi-step creative processes.
The real answers to the questions that come up constantly when people start using Runway.