How To Use Runway
Turn text or a still image into broadcast-quality AI video.
Overview
Runway's Gen-4 turns prompts and reference images into short, cinematic video clips. This guide takes you from your first 5-second clip to a polished sequence you'd actually publish.
Step by step
- 1
Open Gen-4 in the Runway app
From the dashboard, choose 'Generate Video' → Gen-4.
- 2
Start image-to-video, not text-to-video
Generate a hero still in Midjourney first, upload it, then let Runway animate it. Massively better consistency than pure text-to-video.
- 3
Write a motion prompt
Describe what should happen, not what's in the frame. Example: 'slow dolly in, hair lifts in the wind, soft rim light from the right'.
- 4
Pick duration and aspect ratio
5s or 10s. Use 16:9 for YouTube, 9:16 for Reels/Shorts/TikTok.
- 5
Extend or stitch clips
Use 'Extend' on a great clip to keep motion going, or stitch multiple clips in the Runway timeline.
- 6
Add voice and music
Drop in ElevenLabs voiceover and a music bed — Runway can mix natively.
Tips
- Keep camera moves small. Subtle dolly + parallax looks pro; big moves break consistency.
- Reference images with strong lighting direction animate way better than flat ones.
- Generate 4 takes of every clip — costs little, saves hours.
Common mistakes
- • Asking for 5 things in one motion prompt — pick one main motion per clip.
- • Using busy, multi-subject reference frames — Runway gets confused.
- • Forgetting to set aspect ratio before generating — re-rolls cost credits.