en en
HomeRunway MLFeaturesText to Video
Feature

Text to Video

Generate cinematic video clips from text descriptions alone

Overview

Text to Video is Runway's most accessible and powerful generation mode. Describe a scene, motion, camera movement, and mood in plain language, and Gen-3 Alpha renders a high-quality video clip in seconds. With the right prompting technique, results rival footage shot with professional camera equipment — making it the most democratizing tool in video production history.

Gen-3 Alpha produces the highest quality text-to-video results available
Supports complex scene descriptions with multiple elements and interactions
Camera movement descriptors produce genuinely cinematic results
Generates up to 10 seconds of video per clip
Consistent lighting, physics, and temporal coherence across the clip
Works across all genres: cinematic, abstract, nature, product, character

How It Works

1

Write Your Scene Description

Describe the subject, environment, lighting, mood, and most importantly — the camera movement. Include specific cinematic language: "slow dolly forward", "aerial crane rising", "gentle pan left".

2

Select Your Model

Choose Gen-3 Alpha for the highest quality results, or Gen-3 Alpha Turbo for faster generation at lower credit cost. Gen-2 is available on the free plan for experimentation.

3

Set Duration and Aspect Ratio

Choose clip duration (up to 10 seconds) and aspect ratio (16:9 for landscape, 9:16 for vertical/mobile, 1:1 for square). Match your intended platform.

4

Generate and Select

Runway generates your clip. Review it and either use it directly, generate variations with slightly different prompts, or extend it using the Extend feature to add more seconds.

Real-World Examples

Brand Film Opening

Agency creating a cinematic brand film opener

Aerial shot slowly descending through morning mist over a pristine mountain lake, golden hour light reflecting on the water surface, cinematic, peaceful and majestic atmosphere, slow camera descent

Product Reveal

E-commerce brand creating a product launch video

A sleek black smartwatch emerging from darkness, dramatic spotlight illuminating it from above, slow rotation, luxury product reveal, cinematic close-up, dark background with subtle reflections

Abstract Intro

YouTube creator needing a channel intro

Abstract digital particles forming a glowing sphere, deep space background, electric blue and white energy, particles accelerating and converging, dramatic and futuristic, slow motion

Pro Tips

Camera Language is Everything

The single biggest improvement you can make is adding specific camera movement: "slow push in", "gentle pan right", "aerial crane rising", "handheld tracking shot". This transforms generic clips into cinematic footage.

Describe the End State

Tell Runway what the scene looks like at the end of the clip, not just the beginning. "A flower blooming from bud to full bloom" gives the model a clear motion arc to follow.

Use Cinematic Reference Language

Reference film techniques: "shallow depth of field", "anamorphic lens flare", "film grain", "rack focus from foreground to background". This vocabulary produces more professional-looking results.

Generate Multiple Variations

Always generate 3-4 variations of each clip. AI video generation is probabilistic — the same prompt can produce very different results. Budget for multiple attempts and select the best.

Watch Out For

  • Hands and faces can be inconsistent in close-up shots — use medium or wide shots for characters to avoid anatomical issues.
  • Very complex scenes with many simultaneous actions can produce chaotic results — focus each clip on one clear motion or action.
  • Text within video is unreliable — add text overlays in post-production rather than trying to generate them.
Back to Runway ML

You have29:55of free access remaining

Ready to start? Start with ChatGPT basics

Get Started
Chat with us on WhatsApp