How to prompt Runway Gen-4 Image
Runway · Runway web · Guidance v2, verified Sep 23, 2026 against 3 official sources
How to structure your prompt
A descriptive visual prompt, not a conversational or command-style request. Full natural-language sentences give the most control; keywords appended at the end add variation. Cover only what matters among subject, scene, composition, lighting, color, style, focus, angle, text and mood, using concrete visual terms instead of abstract words such as 'beautiful' or 'professional'. Phrase everything positively ('a bald man', not 'a man with no hair'). Set aspect ratio, quantity, resolution and aesthetic range in Runway Tool mode.
Length and syntax
Up to 1,000 characters. Longer prompts are not better if the extra words do not change what appears on screen.
In-image text is described as part of an object, with the exact word written out (for example 'a neon sign that spells OPEN'). Runway's own examples write the word in capitals without quotation marks. No parameter flags.
Prompt skeleton
[Style or medium] of [subject with specific details] in [scene]. [Composition and angle]. [Lighting and color]. [Focus]. [Mood]. [Optional appended style keywords].
What it’s best at
- photographic style control
- compositions and lighting
- visual aesthetic exploration
- film, art and texture styles from keyword collections
- short words on signs and objects
Known limitations
- negative prompts are not supported and can produce the opposite result
- 4:5 is not among the documented native aspect ratios
- image references require a separate workflow
Avoid in your prompt
- negative phrasing
- conversational requests
- unsupported aspect ratio
- abstract adjectives without visual detail
- conflicting instructions (for example dramatic shadows with soft, even lighting)
Other image models
Summarised from the provider’s official documentation by Promptuned. Results vary by model version and settings. Runway Gen-4 Image is a trademark of its owner; Promptuned is not affiliated with Runway.