AI Prompts for Runway ML: 80+ Text‑to‑Video & Image Ideas (Gen‑3/Gen‑4 Guide 2026)

How Runway ML Prompts Work

Runway ML's Gen-series models (Gen-2, Gen-3 Alpha, Gen-3 Alpha Turbo, and Gen-4) generate short, cinematic videos and high-end images from text, images, or video clips. Runway's own and third-party guides suggest splitting your prompt into camera, scene, and subject so the model knows how to frame and move the shot.

Examples of camera/movement keywords include "static close-up shot", "slow dolly-in", "wide aerial shot", "handheld walk-through", "FPV drone shot", and "slow pan across". Runway also encourages descriptive, non-conversational prompts of 15–50 words and positive phrasing rather than "don't do X".

Text‑to‑Video Prompt Formula & Tips

For text-to-video, Runway resources suggest:

• Start with a simple subject + setting, then add motion, lighting, and mood as you iterate.
• Use verbs to describe movement: walking, running, drifting, zooming, panning, tilting.
• Place camera instructions first, then describe the scene and action.
• Use style modifiers like "cinematic", "documentary", "stylized animation", "stop-motion", "90s camcorder", "anime", etc.

30+ Cinematic & Storytelling Prompts (Text‑to‑Video)

These are ideal for short cinematic clips, B-roll, and concept sequences:

• Slow dolly-in shot: a lone figure standing on a cliff above a futuristic city at sunset. Neon signs flicker below, clouds glowing orange, cinematic, moody atmosphere, high detail.
• Static wide shot: quiet street in Paris at dawn. A cyclist rides past cafés, warm window lights, gentle morning fog, soft pastel color grade, filmic look.
• Handheld medium shot: a young woman walking through a Tokyo alleyway at night, surrounded by neon signs and light rain, reflections on wet pavement, cinematic cyberpunk style.
• Wide aerial shot: waves crashing against dramatic cliffs on a stormy coastline, dark clouds overhead, occasional lightning flashes, high-contrast, documentary style.
• Slow pan across: a bustling open-plan startup office, people collaborating at desks, screens glowing, warm practical lighting, subtle camera shake, realistic.
• Low-angle static shot: skyscrapers towering overhead in New York, clouds moving fast between buildings, time-lapse feeling, cool blue tone.
• FPV drone shot: flying through a dense forest in autumn, orange and red leaves, sunlight flickering through branches, immersive, slightly shaky.
• Close-up shot: coffee being poured into a ceramic mug, steam rising, shallow depth of field, cozy kitchen background, warm and inviting lighting.
• Tracking shot: a runner jogging along a riverside path at sunrise, golden light, city skyline in the distance, motivational mood.
• Static medium shot: a person typing on a laptop at a wooden desk, light from the screen reflecting on their face, background softly out of focus, late-night work vibe.

20+ Product, Brand & Social Video Prompts

These prompts suit product teasers, ads, and social B-roll:

• Static product shot: a premium smartwatch rotating slowly on a reflective black surface, soft studio lighting, subtle reflections, clean background, cinematic product style.
• Slow dolly-in: a pair of running shoes on a gym floor, dust motes in the air, spotlight from above, epic sports vibe.
• Static overhead shot: a flat-lay of a laptop, notebook, phone, and coffee cup on a wooden desk, camera slowly drifting slightly, cozy productivity scene.
• Slow pan: a row of skincare products on a marble shelf, sunlight streaming through a bathroom window, soft lens flares, fresh and clean.
• Static medium shot: a barista pouring latte art in a café, focus on the cup, background bokeh, warm golden lighting.
• Tracking shot: a person unboxing a product on a table, tearing packaging, lifting the item, natural home light, vlog style.
• Vertical tracking shot: a person scrolling on their phone in a café, focus on hands and screen, social media feed visible, cozy environment.
• Static wide shot: yoga class in a studio, people moving through poses in sync, sunbeams from windows, calm and peaceful.
• Handheld shot: behind-the-scenes in a kitchen, chef plating dishes, steam rising, energetic documentary style.
• Horizontal pan: city skyline time-lapse, clouds moving fast, day turning to dusk, cinematic tone.

20+ Stylized & Experimental Video Prompts

These prompts let Runway lean into artistic and experimental looks:

• Static wide shot: a city street gradually transforming from realistic to low-poly 3D, colors becoming flat and simplified, surreal art direction.
• Slow dolly-in: a person standing in a room where gravity is flipped, furniture on the ceiling, dreamlike pastel color palette.
• Static shot: a forest where trees are made of glass, light refracting through branches, slow camera push, ethereal mood.
• Top-down shot: labyrinth made of neon lines, glowing pathways, camera slowly drifting above, cyberpunk aesthetic.
• Slow zoom out: a cup of coffee that reveals an entire galaxy swirling inside, stars and nebulae moving, vibrant cosmic colors.
• Static wide shot: an endless hallway of doors floating in space, each door slightly opening as camera moves, mysterious tone.
• Static shot: an urban street where all cars are replaced by floating glowing cubes, night scene, neon reflections.
• Wide shot: a desert made of folded paper dunes, origami cactus, soft shadows, stop-motion inspired look.
• Slow pan: collage wall of vintage photos and newspaper clippings, some elements gently animating, nostalgic atmosphere.
• Static shot: a person dissolving into a cloud of butterflies that fly away, magical realism, warm lighting.

Best Practices for Runway ML Prompts

Across Runway docs and expert guides, a few patterns show up repeatedly:

• Lead with camera: Start prompts with camera type and movement for video—"low-angle static shot", "slow dolly-in", "wide aerial shot".
• Separate sentences: Use one sentence for camera shot, one for scene, one for subject's action, and one for mood/details to reduce confusion.
• Use action verbs: Describe movement clearly ("walking", "panning across", "waves crashing") rather than just static scenes.
• Keep length moderate: Prompts of roughly 15–25 words often balance detail and clarity best for relevance.
• Iterate with previews: Generate short previews, then refine prompts by adding or removing details, and test style presets before committing credits.