0.1% Expert Guide · 2026 Edition

The Complete
Runway AI
Creator Handbook

Everything you need — models, prompting, pricing, camera moves, aspect ratios, workflows, and a newbie start guide — all in one place.

Gen-4.5Flagship Model
6+Core Tools
25cr/sGen-4.5 Cost
$12/moEntry Paid Plan
🚀 Start Here 🤖 Models ✨ Features 💳 Pricing ✍️ Prompting 🎬 Camera Terms 📐 Aspect Ratios ⚙️ Workflows 🎯 Use Cases 💡 Pro Tips ❓ FAQ
Overview

What is Runway?

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.

🎥

AI Video Generation

Text-to-video and image-to-video using Gen-4.5, the world's top-rated video generation model in 2026.

✂️

Video Editing (Aleph)

Edit existing video with text prompts — swap backgrounds, add objects, change lighting, generate new angles.

🎭

Motion Capture (Act-Two)

Transfer facial expressions, body movement, and performance from any video to a digital character. No suit needed.

🔊

Generative Audio

Add speech, lip-sync, and audio to still images and videos using AI-generated voice tools.

🤖

General World Models

GWM-1 powers real-time simulation, explorable environments (Worlds), and conversational video avatars (Characters).

🔧

Custom Workflows

Build automated, multi-step AI pipelines for character generation, storyboarding, and creative production.

New to Runway

Newbie Starter Guide

Never used Runway before? Start here. This is the exact path from zero to your first real video generation in under 30 minutes.

ℹ️
Before you start: The Free plan gives you 125 one-time credits (never refreshes). That's enough for 5 seconds of Gen-4.5 video or 25 seconds of Gen-4 Turbo. Use it to explore the interface, then upgrade to Standard ($12/mo) to actually work.
01

Create Your Free Account

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.

02

Understand the Dashboard

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.

03

Pick Your First Mode: Text to Video

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.

First Prompt to Try

"A slow dolly shot through a misty forest at golden hour. Sunbeams filter through tall pine trees. Cinematic color grading. Shallow depth of field."

04

Try Image to Video

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.

05

Learn the Iteration Loop

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.

06

Explore the Runway Academy

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.

07

Decide on a Plan

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).

AI Models

Runway's Model Suite

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.

Gen-4.5
Flagship

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.

Credit Cost
25 cr/sec
Duration
5 or 10 sec
Modes
T2V + I2V
Min Plan
Standard
Gen Time
~60–90 sec
Max Export
4K
Gen-4
Solid Quality

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.

Credit Cost
12 cr/sec
Duration
5 or 10 sec
Modes
T2V + I2V
Min Plan
Standard
References
Yes
Gen-4 Turbo
Fast & Cheap

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.

Credit Cost
5 cr/sec
Duration
5 or 10 sec
Best Use
Iteration
Explore Mode
Unlimited ✓
Aleph 2.0
NEW

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.

Min Plan
Standard
Input
Video + Text
Version
Aleph 2.0
Green Screen
Yes
Act-Two
Motion Capture

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.

Tracking
Face + Body + Hands
Voice
Change Voice ✓
Input
Any camera
Min Plan
Standard
GWM-1 (General World Model)
Research

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).

Variants
Worlds / Avatars / Robotics
Real-time
Yes
Characters API
Available
🔗
Multi-Model Dashboard: All paid plans include access to third-party models built right into Runway — Kling 3.0 Pro, Veo 3 and 3.1, Seedance 2.0, BFL FLUX.2 [max], and Seedream 5.0. You no longer need four separate subscriptions. Pick the right model for the job from a single interface.
Platform Features

All Features Explained

Runway is more than a video generator. Here's the full feature set broken down by what it does and how to access it.

📝 Core

Text to Video

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.

🖼️ Core

Image to Video

Upload an image as the first frame, then write a motion prompt to animate it. Your most powerful mode for control and visual consistency.

🎯 Gen-4+

References System

Upload reference images to lock in specific characters, locations, or objects across multiple generations. Essential for narrative and multi-shot work.

🔄 Aleph 2.0

Video Transformation

Edit real or generated video with text prompts. Change lighting, style, backgrounds, and scene elements without regenerating the whole clip.

📐 Aleph 2.0

New Angle Generation

Generate different camera angles from a single source shot. Create wider or tighter framings from existing footage — useful for coverage you didn't capture.

🟢 Aleph 2.0

Green Screen Creator

Turn any video into a green screen asset or clean plate. No rotoscoping required. Eliminates hours of manual masking work.

🎪 Act-Two

Performance Transfer

Film yourself or anyone performing, then transfer that performance — face, body, hands — to a digital character of any style.

🗣️ Audio

Generative Audio & Lip Sync

Generate speech and add lip-sync to still portraits and videos. Custom voice creation available on Pro tier and above.

⬆️ Standard+

Upscaling

Upscale generated videos to higher resolution for final delivery. Available on all paid plans with watermark removal included.

💬 Gen-4+

Chat Mode

A conversational generation interface — generate anything using Gen-4 images, videos, and references from within a single chat UI. Available to all users.

🎞️ Editor

Video Editor (Edit Studio)

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.

🔧 Advanced

Custom Workflows

Build automated multi-step AI pipelines. Featured workflows include Character Creator, Storyboard Creator, and Story Panels. Advanced users can build fully custom pipelines.

🌐 API

Runway API

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.

🧑‍🎤 GWM-1

Runway Characters

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.

🏋️ Enterprise

Custom Model Training

Train custom models on your own visual style, brand, or characters. Used by studios like Lionsgate for bespoke AI production pipelines.

Plans & Credits

Pricing & Credit System

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.

Free
$0
125 one-time credits (never refresh)
  • Gen-4 Turbo (image-to-video only)
  • Text-to-image
  • Text-to-speech & audio apps
  • 3 video editor projects
  • 5 GB storage
  • Watermark on all exports
  • No Gen-4 or Gen-4.5 video
Unlimited
$76/mo (annual)
2,250 credits + Explore Mode
  • Everything in Pro
  • Explore Mode: unlimited slow-queue generations of Gen-4 Turbo, Aleph, Act-Two, Gen-3 Alpha Turbo
  • Gen-4.5 still draws from credits
  • Best for studios & heavy iteration
Enterprise
Custom
Custom credits + SLAs
  • Custom model training
  • Dedicated support
  • API priority access
  • Lionsgate, NVIDIA partnerships
  • Education discounts available
⚠️
Credit Reality Check: 625 credits (Standard) = ~25 seconds of Gen-4.5 video total for the month. That's approximately 5 clips at 5 seconds each. Budget carefully. Use Gen-4 Turbo for iteration and Gen-4.5 for your final shots only.

Credit Cost Per Model

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
Effective Prompting

The Complete Prompting Guide

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.

Core Prompt Formulas

Text to Video — Official Structure
[Camera movement] shot of [subject/object] [action] in [environment]. [Supporting details].
Example

"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."

Include: visual description + motion description. Both are required for strong results.
Image to Video — Official Structure
The camera [motion description] as the subject [action]. [Additional motion details].
Example

"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."

Key rule for I2V: your image defines the visual. Your prompt defines the motion. Don't re-describe what's in the image — describe what should happen.
Sequential Prompting (Timestamps)
[00:01] Subject stands still. [00:03] Camera pushes in slowly. [00:05] Subject turns to face camera, smiling.
Use timestamps for precise temporal control. Make sure the requested sequences fit within your chosen duration (5 or 10 seconds).

Ready-to-Use Prompt Templates

Cinematic Scene

"A [shot type] of [subject] in [setting]. The camera [movement type] as [subject motion]. [Lighting description]. [Style/mood]. [Additional environmental details]."

Product Ad

"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."

Social Content (Vertical)

"Handheld vertical shot. [Subject] doing [action] in [location]. Natural lighting. Authentic, unpolished energy. Documentary feel. Shot on phone. [Any specific motion detail]."

Image to Video (Motion Focus)

"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."

Prompt Components — Visual vs. Motion

👁️

Visual Components

  • Subject appearance
  • Environment / setting
  • Lighting setup
  • Framing / composition
  • Style / aesthetic
🎬

Motion Components

  • Subject action
  • Environmental motion
  • Camera movement
  • Motion style & timing
  • Direction & speed

✅ Do This

  • Use positive language — describe what you WANT, not what you don't want
  • Start simple, then add one component at a time during iteration
  • Use concrete, observable descriptions ("neon signs reflect on wet pavement")
  • Name specific camera moves ("slow dolly push," "aggressive whip pan")
  • Include motion timing cues when needed ("slowly at first, then quickens")
  • Reference cinematography language — DP vocabulary works well
  • For I2V: focus almost entirely on motion, not visual description
  • Add "Continuous, seamless shot" when cuts appear unexpectedly

❌ Avoid This

  • Vague emotional language ("a beautiful, mysterious feeling")
  • Conflicting instructions in the same prompt
  • Multi-paragraph prompts that over-specify every detail
  • Re-describing what's in the image when doing I2V
  • Conversational phrasing ("Please make a video of...")
  • Negative prompts ("no motion blur") — describe what you want instead
  • Concepts without visuals ("a sense of longing and despair")
  • Expecting perfection on the first generation
💡
Iteration Strategy: Extremely complex, multi-paragraph prompts paradoxically lead to worse results — the model struggles to honor every detail simultaneously. Start simple, get a baseline, then add one specific element per iteration. Each generation teaches you how the model interprets your language.
Camera Language

Camera Movements & Techniques

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.

Movement Types

Dolly Push / Pull
Camera physically moves toward or away from subject. More cinematic than zoom.
Pan
Camera rotates horizontally on a fixed axis. Good for following subjects or revealing environments.
Tilt
Camera rotates vertically. Use to reveal height or emphasize scale (tilting up a building).
Orbit / Arc
Camera circles around the subject. "Sweeping horizontal arc" creates dramatic 3D reveals.
Tracking Shot
Camera follows subject from the side while subject moves. Creates kinetic energy.
Whip Pan
Ultra-fast horizontal pan creating a blur effect. Used for cuts between subjects or high-energy transitions.
Zoom
Optical zoom in/out (less cinematic than dolly). "Crash zoom" is a rapid, aggressive zoom-in.
Handheld
Natural camera shake, imperfect movement. Adds documentary/authentic energy. "Extreme shakiness" for chaos.
Steadicam
Stabilized handheld — smooth gliding movement while walking. Follows subject fluidly.
Locked Off
Camera completely still. "The locked-off camera remains perfectly still." Use for static scene motion.
FPV (First Person)
First-person view as if through the subject's eyes or a drone. Immersive, disorienting.
Low Angle
Camera positioned below eye level, angled upward. Creates power, scale, and drama.
High Angle
Camera above the subject looking down. Vulnerability, overview, or God's-eye perspective.
Bird's Eye / Drone
Top-down aerial view. "Overhead drone shot slowly pulling back."
Dutch Angle
Tilted horizon line. Creates psychological unease, tension, or disorientation.
Rack Focus
Shift of focus from one element to another within the frame. "Focus shifts from foreground to background."

Combining Camera Terms

Combining terms is encouraged — especially for cinematic sequences. Mix camera angles, motion style, and composition freely.

Combined Example

"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."

Output Formats

Aspect Ratios & Output Specs

Runway supports multiple aspect ratios depending on the mode and model you're using. Choose your ratio based on where the content will live.

16:9
Widescreen / Landscape. YouTube, film, presentations, TV. Default for text-to-video.
9:16
Vertical / Portrait. TikTok, Instagram Reels, YouTube Shorts. Supported natively at 720×1280 in I2V.
1:1
Square. Instagram feed posts, display ads, some social platforms.
📋
Duration Options: 5 seconds or 10 seconds per clip. For longer sequences, extract the last frame of one generation and use it as the input image for the next — then combine clips in a video editor. This is the professional workflow for building multi-shot sequences.
⚠️
Text to Video note: As of early 2026, Text to Video in Gen-4.5 is locked to 16:9. The 9:16 vertical format is available natively via Image to Video. Verify current support at app.runwayml.com as Runway updates frequently.
Production Workflows

Workflows & Automation

Runway Workflows let you build multi-step AI pipelines — automating complex creative processes and creating consistent, repeatable results at scale.

🧑‍🎨 Featured

Character Creator

Generate consistent characters across multiple shots and styles. Uses References + Workflows to maintain visual coherence.

📋 Featured

Storyboard Creator

Build full visual storyboards from a concept. Generate sequential story panels with consistent world and character.

🌍 Featured

Story Panels

Build an entire world from a single reference image. Create consistent sequences across multiple shots using one source.

⚙️ Custom

Custom Pipeline Builder

Build fully custom automated workflows for your production needs. Advanced users can chain models, set conditions, and automate entire production stages.

Professional Production Workflow (Recommended)

Phase 1
Iterate fast with Gen-4 Turbo

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.

Phase 2
Finalize with Gen-4.5

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.

Phase 3
Edit and polish with Aleph

Use Aleph 2.0 to refine the generated footage — adjust lighting, swap backgrounds, clean up problem areas, or generate coverage shots from existing clips.

Phase 4
Extend sequences using last-frame technique

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.

Phase 5
Add performance with Act-Two

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.

Who It's For

Use Cases by Industry

Runway is used across industries by professionals who need cinematic AI video in their workflow. Here's how each sector uses it.

🎬

Media & Entertainment

Pre-visualization, VFX shot development, short film production, concept pitching. Used by studios in partnership with Lionsgate and others.

📢

Marketing & Advertising

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.

🎮

Gaming & Interactive

Cinematic cutscenes, concept art animation, character movement, world-building assets. AI for Games course available free in Academy.

💡

Product & Design

Product visualization, concept exploration, UI/UX motion references, design presentation videos.

🏛️

Architecture & Design

Animated architectural walkthroughs from static renders. KPF uses Runway to animate projects in-house, replacing expensive external rendering pipelines.

👗

Fashion

Lookbook content, garment motion studies, editorial content, runway (literal) simulations. Generate diverse model representations with consistent styling.

🎓

Education

UCLA Film, TV and Digital Media uses Runway to empower students to experiment with AI in filmmaking. Academic pricing available.

📣

Agencies & Professional Services

Client presentation content, pitch decks, rapid ideation and concepting, multi-channel campaign asset creation at scale.

Expert Knowledge

Pro Tips from Power Users

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.

TIP 01
Always check your image for implied motion before doing I2V

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.

TIP 02
Use the 5-10 second duration strategically

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.

TIP 03
The last-frame method is how professionals make long sequences

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.

TIP 04
Think like a Director of Photography, not a writer

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.

TIP 05
Use Gen-4 References for any multi-shot project

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.

TIP 06
Aleph is the most underused tool on the platform

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.

TIP 07
Over-specification paradox is real

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.

TIP 08
Budget math: iterate on Turbo, finalize on 4.5

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.

TIP 09
Known limitations to work around in 2026

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.

TIP 10
Free Academy courses are legitimately excellent

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.

Learning Resources

Runway Academy — Free Courses

Academy.runwayml.com hosts 31 courses and tutorials — all free. Here are the most valuable ones and who they're for.

📣 Beginner

AI for Advertising

10 modules. Product shots, social ads, campaign content. Start here if your goal is commercial content creation.

🎥 Intermediate

AI for Visual Effects

6 modules. Integrate AI into VFX pipelines, pre-visualization, and production-ready shots.

🎮 Intermediate

AI for Games

5 modules. Asset creation, cinematics, and game production pipelines with Runway tools.

🔄 Intermediate

Video Transformation with Aleph

5 modules. Add scene elements, change environments, generate new angles, transform videos with natural language.

🎭 Intermediate

Character Animation with Act-Two

2 modules. Capture facial expressions, body gestures, and change character voices for dynamic animation.

⚙️ Advanced

Building Custom Workflows

3 modules. Build automated AI pipelines for character generation, storyboarding, and multi-step creative processes.

FAQ

Frequently Asked Questions

The real answers to the questions that come up constantly when people start using Runway.

The free plan gives you 125 one-time credits that never refresh — that's roughly 5 seconds of Gen-4.5 video or 25 seconds of Gen-4 Turbo. It's enough to see what the platform does, but not enough to actually work in it. Standard at $12/month is the real entry point. It unlocks Gen-4.5, Aleph, Act-Two, watermark removal, and all third-party models.
Text to Video builds everything from a written description — composition, subject, lighting, environment, and motion. Image to Video takes an uploaded image as the first frame (defining composition, style, and visual) and uses your text prompt exclusively to direct motion. Image to Video gives you significantly more visual control because you provide the starting visual directly. Most production workflows favor I2V for consistency.
Cuts usually mean your image/prompt combination needs more duration. Try increasing from 5 to 10 seconds first. If cuts continue, check your prompt for any phrasing that implies a scene change, and add "Continuous, seamless shot" explicitly to your prompt. You can also add stabilization in a video editor after the fact.
Use the Gen-4 References system. Upload a clear, well-lit reference image of your character and use the @ recall feature in your prompts. The updated version of Gen-4 References significantly improves object and character consistency. For even more control, use the Character Creator featured Workflow in Workflows. This is the essential step for any narrative or multi-shot project.
Use the last-frame technique: scrub your completed video to the very last frame, click "Use," then select "Use current frame." This loads it as the input image for a new generation. The model matches the visual style and motion direction, creating a seamless continuation. Combine the clips in Edit Studio (Runway's built-in video editor) and trim the shared frame.
They're different tools for different jobs. Gen-4.5 wins blind preference benchmarks for shot-level visual fidelity and prompt adherence to camera direction. Veo 3.1 wins on native audio generation and lip-sync. Kling 3.0 wins on 4K resolution and multi-shot duration (up to 2 minutes). Seedance 2.0 wins on multimodal references in a single pass. The smart answer: use the right tool for the job. Conveniently, all paid Runway plans now include access to all these models in one dashboard — you don't need separate subscriptions.
Explore Mode gives you unlimited generations of Gen-4 Turbo, Aleph, Act-Two, and Gen-3 Alpha Turbo at a relaxed (lower-priority) processing speed — typically 10-20 minute queue times. Critically, Gen-4.5 is NOT included in Explore Mode — it still draws from your 2,250 monthly credits. Explore Mode is most valuable for studios and heavy users who iterate constantly and want to remove the mental friction of credit counting during the draft phase.
Most likely because the prompt is describing a feeling rather than a shot. Runway reads camera direction (lens height, movement, timing), subject action (gait, gesture), and lighting geometry — not emotional intent. Replace "mysterious and dark energy with slow suspense" with "slow dolly push at knee height, shallow depth of field, low-key lighting with a single rim light." Think like a DP, not a writer. Also check your input image for implied motion cues that may be contradicting your prompt.
Known limitations as of mid-2026: text rendering within video remains inconsistent (signs and labels often contain errors); very specific product details may not match real products without image references; human hands and fine motor actions can still produce artifacts; 16-second max duration per clip (versus Kling's 2 minutes); no native audio generation (unlike Veo 3.1). For pixel-perfect product accuracy, always use image-to-video with real product photography rather than text-to-video alone.
1) Take the free Runway Academy course that matches your use case (Advertising for marketers, VFX for filmmakers, Games for game devs). 2) Read the full Prompting Guide at academy.runwayml.com/guides/prompting-guide. 3) Generate one simple clip, review the output, and change one thing. Repeat 10 times. The iteration loop teaches you more than any tutorial. 4) Watch the "Behind the Scenes" tutorials to see how real creators actually build sequences. 5) Join the Discord community (discord.gg/runwayml) to see what prompts other creators are sharing.