Runway Video Prompts: Shot List Language That Works

Runway video prompts work when each clip reads like one row on a shot list: framing, angle, camera move, subject action, lens feel, light, and a clear duration. Gen-4 fills gaps when you write vibe adjectives instead of production fields. This guide teaches shot list language you paste into Runway text-to-video and image-to-video boxes as of August 2026. You leave with a field order, conversion steps from spreadsheet row to paste, multi-shot continuity habits, and a soft path to PromptMake /image when a hero still locks the subject before motion. For the SMCD spine across hosts, see the ai video prompt structure guide. For Gen-4 feature dialect, see the Gen-4 prompts guide when it ships beside this page.

What shot list language means for Runway video prompts

Shot list language names production fields in a fixed order so the model hears one decision per clause. A film shot list tells a crew what to shoot. A Runway shot list tells Gen-4 what to invent for five to ten seconds. Same vocabulary: wide, medium, close-up, eye-level, low angle, slow dolly in, pan left, locked tripod. Different job: you convert rows into sentences the text box accepts.

You need this craft if you cut product teasers, social B-roll, concept pre-viz, or short narrative beats inside Runway. You also need it if a spreadsheet or AI shot planner outputs columns you refuse to paste as raw tables. Columns become clauses. Clauses become one paste per generation pass.

This page extends video basics. Subject, motion, camera, and duration still decide whether a clip holds. Shot list language is how you write those pillars in film terms Runway already rewards. It is not a Gen-4 feature tour, not a pan-dolly-orbit keyword encyclopedia, and not a fifteen-line paste bank for blank-canvas starters. Those jobs live on sibling guides. Stay here when your brief starts as rows: shot ID, framing, move, action, seconds.

Skip deep shot list work if you only export still frames from Midjourney or FLUX and never open a motion UI. Lean in if you burn credits on "cinematic masterpiece" lines that invent three camera paths in four seconds.

The shot list fields that become prompt clauses

Every useful Runway row maps to a prompt clause. Write the fields in production order even when you merge them into one paragraph later. Order trains your eye. Merge trains the paste box. As of mid-2026, Runway Gen-4 responds to named cinematography terms more than mood adjectives. "Slow dolly push-in over five seconds" beats "dynamic cinematic energy."

Keep one primary subject action and one camera path per row. Split montage ideas across rows, then generate separate clips and edit. A single Runway pass that tries to play five beats collapses into mush. Treat duration as a hard budget for those two decisions.

Positive phrasing matters on Gen-4. Describe what happens. "Locked tripod, steady frame" lands. "No camera shake" can pull the model toward shake. Build each clause as a yes statement. The subsections below define the fields and give paste fragments you can chain.

Framing, angle, and shot size

Shot size sets how much of the subject fills the frame: extreme wide, wide, medium, medium close-up, close-up, extreme close-up, insert. Angle sets viewpoint: eye-level, low angle, high angle, overhead, POV, slight Dutch tilt when unease is the point. Lead runway video prompts with these words so Gen-4 locks composition before it invents motion.

Paste fragments: "Wide eye-level shot," "Medium low-angle shot," "Close-up eye-level," "Overhead insert of the watch face." Pair size and angle in the first clause. Example open: "Medium eye-level shot of a brushed steel bottle on gray stone." Weak open: "Beautiful product hero, cinematic."

Vertical social lanes (9:16) shrink horizontal travel. Prefer medium and close sizes on 9:16 when labels or faces must stay readable. Save extreme wides for 16:9 establishes where side space exists for real.

Camera move, speed, and subject action

Camera move names how the lens travels: locked tripod, slow pan left, gentle tilt up, slow dolly in, lateral truck right, quarter-orbit clockwise. Speed words that travel: slow, gentle, measured. Speed words that break short clips: whip, crash, frantic. Subject action names what the actor or prop does inside those seconds: steam curls once, lid opens, walker takes three steps, brake lights brighten.

Separate the two systems in text. "Bottle static; camera slow dolly in only" keeps identity. "Dynamic reveal" merges them and both fail. Time the action to the UI seconds: "0–2s static, 2–5s lid opens once." Runway still reads that spine when you merge into prose.

One move per row. Pan plus dolly plus orbit in a five-second box is a credit sink. If the brief needs two moves, write two shot IDs and generate twice.

Lens, light, duration, and what the shot reveals

Lens feel anchors aesthetic: 24mm wide, 35mm spherical, 50mm, 85mm product feel, shallow depth on subject. Lighting sets mood without eating motion budget: soft key from upper left, overcast daylight, neon practicals, tungsten practicals. Duration belongs in the Runway slider and once in the text when it clarifies pacing: "over five seconds," "six-second hold."

Add a short reveal clause when the shot has a narrative job: "The push-in builds quiet tension," "The pan reveals the window light," "The insert confirms the engraved serial." That clause ties the row to story intent and helps Gen-4 prefer coherent motion over random flourish.

Skip resolution praise, "8K," and "masterpiece" inside the prompt. Set quality in UI controls when they exist. Spend tokens on fields the model can act on.

How to convert a shot list row into a Runway paste

Conversion is a worksheet habit. You fill fields, read them aloud, then merge into one paste. The merge order below matches how Runway camera and video prompting guides describe successful Gen-4 lines as of mid-2026: shot size and angle, camera move with speed, subject and action, lens, lighting, what the shot reveals. Duration sits in the UI and can echo once in prose.

Budget ten minutes for the first conversion. Later rows take two minutes when your template exists. Keep a notes column for host mode (text-to-video vs image-to-video), aspect, and date. That log stops you from blaming the model when the slider sat at four seconds while the beats needed eight.

Image-to-video changes the subject clause. The upload carries pixels. Text carries delta: preserve language, motion, camera. Text-to-video needs full subject nouns in the open. The steps below cover both routes.

Step 1: Fill the row fields before you write prose

1. Shot ID and one-line purpose ("hero push-in," "establish street," "label insert").
2. Shot size + angle.
3. Camera move + speed (or locked).
4. Subject nouns: material, color, count, place.
5. Timed subject action matched to target seconds.
6. Lens feel and light direction.
7. Duration in seconds for the Runway slider.
8. Reveal purpose in one short clause.

Example row for a product hero: ID A1, purpose label read, medium eye-level, slow dolly in only, brushed steel bottle blue cap no logo on gray stone, 0–2s static then 2–5s condensation bead slides once, 85mm product feel soft key upper left, 5s, reveal confirms material finish.

Step 2: Merge into one Runway sentence stack

Merge in field order. Text-to-video paste: "Medium eye-level shot of a brushed steel water bottle with blue cap, no logo, on gray stone. Bottle static two seconds, then one condensation bead slides down the left side. Slow dolly in only over five seconds, 85mm product feel, soft key from upper left. The push-in confirms the brushed finish."

Image-to-video paste: "Maintain subject, lighting, and background from reference. Bottle static two seconds, then one condensation bead slides down the left side. Slow dolly in only over five seconds. Camera path only; wardrobe and label unchanged." Crop the still to the output aspect before upload. A vertical product still forced into a wide lane invents side walls.

Read the merge for duplicate motion verbs. If dolly and orbit both appear, delete one before you spend credits. Confirm the slider matches the seconds you wrote.

Step 3: Generate once, then change one field

Generate a single pass. Score three checks: subject readable, action finished inside the window, camera followed one path. Fix the failing check only. Subject drift: tighten material and color nouns, or lower motion strength if the UI exposes it. Motion mush: cut a beat or add one second. Camera chaos: delete extra move words until one verb remains.

Save the winning paste next to the shot ID. Next week you swap the product noun and keep the move grammar. That habit turns a spreadsheet into a reusable Runway kit without rewriting from scratch.

Multi-shot sequences and continuity language

Most Runway stories need more than one generation. A thirty-second cut is six five-second rows, not one prompt that lists "then," "next," and "after that." Gen-4 and related Runway video lanes favor one action spanning about five to ten seconds per pass. Plan the sequence on paper. Generate clips. Edit in your NLE or timeline tool.

Continuity language keeps character, wardrobe, light, and place stable across rows. Repeat the same subject nouns. Name the same time of day and weather. Keep lens family close when you want a matched look: stay in 35mm and 50mm territory rather than jumping from 14mm to 135mm without a reason.

Last-frame chaining helps when the host supports first or last frame inputs. Export the final frame of clip A. Use it as the start frame for clip B with a new motion clause. That pattern extends a walk, a push-in, or a light change without asking one prompt to cover twelve seconds of montage. Confirm the control name inside your Runway build; public UIs shift through 2026.

Write shot IDs that state the cut intent: A1 establish wide locked, A2 medium push-in, A3 insert of hands. When you paste, each ID is its own generation. Do not paste the full sequence into one box.

Credit discipline: draft all rows on a cheaper or Turbo lane when your account offers one, then rebuild hero shots on the higher Gen-4 quality tier. Keep surrounding B-roll on the draft lane once the flow holds. Exact credit prices change; treat the two-pass habit as the durable rule.

Common mistakes with Runway video prompts

Mistake 1: Vibe stack with no shot fields. "Epic cinematic rain scene, masterpiece, 8K" names no size, move, or timed action.

Mistake 2: Five beats in one five-second row. Split into shot IDs. Generate separate clips.

Mistake 3: Negative camera language. Prefer "locked tripod, steady frame" over "no shake."

Mistake 4: Subject motion and camera motion fused into one vague verb. Split them every time.

Mistake 5: Image caption pasted into a video box. Still language freezes pose. Add move and timed action or accept invent chaos.

Mistake 6: Aspect fight on image-to-video. Crop the hero still to the Runway lane before upload.

Mistake 7: Duration longer than beats. Ten UI seconds with one steam curl wastes credits on empty tail frames.

Mistake 8: Rewriting every field after one miss. Change one axis: framing, move, action, or seconds.

Mistake 9: Treating a shot list table as the paste. Convert columns to clauses first.

Mistake 10: Chasing Gen-4 folklore before the row fields hold. Lock size, move, action, and seconds, then read release notes for dialect tweaks.

Runway notes for mid-2026 (without a Gen-4 feature tour)

Public names inside Runway shift. Confirm the model picker before you paste old forum strings. As of August 2026, Gen-4 remains the main video generation family people mean when they search runway video prompts. Turbo or draft lanes often exist for cheap tests. Higher quality Gen-4 modes suit hero shots once the shot list row is clean.

Shared rules that travel across Runway video modes: lead with shot size and angle, name one camera path with speed, describe subject action in timed beats, use positive phrasing, keep each pass under about ten seconds of action, set duration in the UI to match the beats. Image-to-video text should focus on motion and camera when the upload already locked the subject. Re-describing every pixel from the still fights the reference.

Audio fields appear on some 2026 builds. If your route generates sound, add a short ambient cue or write "no music" after the visual line so a score does not land on a quiet product clip. Keep audio optional. Visual shot fields still decide whether the clip works.

When you need deeper Gen-4 motion strength, reference workflows, and consistency controls, open the dedicated Gen-4 prompts guide on this blog. This page stays on shot list language: how rows become pastes that Gen-4 can obey.

When PromptMake /image fits Runway prep

Runway consumes pixels and text. PromptMake /image shortens the still step before you open a motion lane. Use /image when you need a hero frame that matches your shot list subject nouns, when a noisy phone photo keeps drifting in image-to-video, or when you want Midjourney, FLUX, DALL·E, Stable Diffusion, or Leonardo formatted prompt text from a reference without hand-translating flags.

Workflow: build or select the still for shot A1. Soft start: https://promptmake.net/image. Pick Recreate Exactly to hold composition for an image-to-video upload, Adjust Lighting when exposure fails before motion, Create Variation when you need alternate hero stills that share the same shot list shell. Export a clean PNG. Upload to Runway image-to-video. Paste move and action clauses only.

Guest accounts receive about three /image runs per day. Free registration raises the image quota. Quotas for /image and /text stay separate. Spend image runs on subject lock, not on hunting synonyms for a camera move you have not chosen. PromptMake does not render Runway clips. It prepares the still so your runway video prompts attach to a frame worth animating.

FAQ

What are runway video prompts?

Runway video prompts are written briefs that tell Runway Gen-4 what to generate for a short clip. Strong prompts read like shot list rows: framing, angle, camera move, subject action, lens feel, light, and duration. Weak prompts stack still-image praise words and leave motion vague. Convert each row into one paste, then generate one pass per shot ID.

How do I structure runway video prompts like a shot list?

Fill fields first: shot size and angle, camera move with speed, subject nouns, timed action, lens, light, seconds, and a short reveal purpose. Merge in that order into a sentence stack. Set the Runway duration slider to match the beats. Keep one camera path and one primary action per generation. Split montage ideas across multiple shot IDs.

Do runway video prompts work for image-to-video?

Yes. The upload carries subject pixels. Your text adds preserve language, timed motion, and one camera path. Lead with "Maintain subject, lighting, and background from reference," then write the action and move. Crop the still to the output aspect before upload so Gen-4 does not invent side space.

How long should each Runway shot be?

Plan about five to ten seconds of action per generation on consumer lanes as of August 2026. Match motion scope to the UI slider. One primary action per two to four seconds is a safe starting ratio. Longer stories come from multiple shot IDs edited together, not from one prompt that lists a full montage.

How is this different from a Gen-4 feature guide?

This page teaches shot list language: how production fields become paste clauses. A Gen-4 feature guide covers model-specific motion, camera, and consistency controls in depth. Use both. Learn field order here. Open the Gen-4 guide when you need dialect and control details after your rows already hold.

How do I keep continuity across multiple Runway shots?

Repeat the same subject nouns, wardrobe, time of day, and place across rows. Keep lens family close. Generate separate clips per shot ID. When your build supports first or last frame inputs, chain the last frame of clip A into the start of clip B with a new motion clause. Edit the sequence outside the single-prompt box.

When should I use PromptMake /image with runway video prompts?

Use /image when you want a cleaner hero still or model-ready still prompts from a reference before you spend Runway credits. Soft start: https://promptmake.net/image. After the still holds, either write full subject nouns into a text-to-video paste or upload the still and keep your runway video prompts focused on move and action.