Runway for ads: a practical guide

What Runway does for ad work

Runway is an AI video suite and API: its own models make and edit shots, it resells third-party models, and it has apps and API recipes built for product ads. For ad work the three tools that matter are Gen-4.5 for new shots, Aleph 2.0 for editing footage you already have, and Act-Two for character performance.

Runway describes itself as "building foundational Real-World Intelligence", with a creative suite, an API and a robotics toolkit (Runway). Besides its own models, its API lists Seedance 2.5, Veo 3.1, Gemini Omni Flash, Wan 3.0, MiniMax H3 and others (Runway API models), and its plans list Kling 3.0 (Runway pricing).

Two gaps to plan around. Gen-4.5 outputs 720p, so use the plan's 4K upscaling for final exports. The Gen-4.5 help article does not list audio as an output, so plan the voice and sound separately or use a model that generates audio.

Runway's ad-specific apps and recipes

Runway has tools built for product ads that general video models do not. The API "recipes" take product images and a brief and return a finished clip, and the app has image tools for statics and storyboards.

Product swap is the most useful of these for scaling. If one product demo works, the same shot can carry another product from the same brand without regenerating the scene.

Which ad formats Runway suits

Runway suits formats where the picture carries the ad: product demos, brand film and stylised scenes, plus re-versions of footage that already exists. The table is guidance from published capabilities, not test results.

For a pure talking-head read with many languages, a dedicated avatar tool is often faster. See HeyGen for ads.

Workflow: from script to finished Runway ad

Build the stills first, generate shots from them, then use Aleph 2.0 to make the versions. A 30 second ad on Gen-4.5 is usually four to eight shots of 2 to 10 seconds.

1. Break the script into shots. One line or action per shot, each under 10 seconds for Gen-4.5. Mark which shots are talking shots; those go to Act-Two, Character Script to Video or the product UGC recipe.
2. Make the start frames. Generate or reshoot product and scene stills, then use them as image-to-video inputs. The Gen-4.5 guide says image-to-video prompts "should focus on describing the motion of the scene" (Gen-4.5).
3. Generate in the final ratio. Pick 9:16 for Reels and TikTok at generation time. The Gen-4.5 guide notes that changing the ratio away from the input image's shape crops the input.
4. Record the talking shots. For Act-Two, film yourself or an actor reading the line on a phone, then apply it to the character.
5. Review every take. Check label text, hand and finger count, product shape between frames, and mouth shapes on talking shots.
6. Edit and upscale. Assemble, add captions, voice, music and the end card, then upscale to 4K for delivery. Put prices, offers and claims in as editor text.
7. Version with Aleph 2.0. Change the background, wardrobe, time of day or product in the finished shots to make test variants without regenerating the scene.

Prompt templates for Runway ads

Write Gen-4.5 prompts as a sequence of camera and action instructions, and Aleph 2.0 prompts as one action verb plus the change. Aleph's guide recommends "An action verb that describes what you want to do" and "A description of your desired transformation" (Aleph 2.0 guide).

The templates below are written for this guide. Square-bracket items are placeholders.

Gen-4.5 product hero (image to video, 5 seconds, 9:16)

The camera starts close on the [product] label, then pulls back and rises
to reveal it on [surface] in [setting].
At 3 seconds, [one action: a hand picks it up / the lid opens / light sweeps across].
Soft [time of day] light, shallow depth of field, slow and steady movement.

Gen-4.5 brand-film establishing shot (text to video, 10 seconds)

Wide shot of [location] at [time of day], [weather].
A [person or vehicle] enters from frame left and crosses to the centre.
The camera tracks right at walking pace, then holds as they stop.
Muted [palette] colours, light film grain, no text.

Aleph 2.0 version edits (one per generation)

Change the background to [new setting], keep the person and product unchanged.
Change the outfit to [colour and garment].
Re-light the scene as [time of day] with warm light from [direction].
Replace the [original object] with [product], label facing camera.

Product UGC recipe creative direction (userConcept field)

Tone: casual, first person, talking to a friend.
Hook: "[hook line]"
Script: "[problem in one line]. I started using [product]. [One specific result the brand can document]. [Offer]."
Setting: [car / kitchen / bathroom mirror], natural light, handheld.

Act-Two driving performance checklist (what to film)

Frame: shoulders up, face fully in shot, plain background.
Light: even, from the front.
Read: [line], once at normal pace, once slower.
Gestures: [one hand gesture] on the word "[key word]".
Audio: quiet room, phone close to the mouth.

Limits, terms and policy notes

Runway's terms allow commercial use on every plan, but free videos carry a watermark and outputs may be used for model training. Check both before you upload brand assets.

Commercial use. The Terms of Use (4.4) say: "The Company does not claim ownership of any of your Inputs or Outputs" and "Subject to your compliance with the Agreement, the Company does not restrict your commercial use of your Outputs" (Runway terms). The help center lists "product advertising" as a covered use and says there is "no formal requirement for you to credit us" (Usage rights).

Training. Section 4.4 also says "Inputs and Outputs may be used by the Company to train and improve its AI models." Ask the brand before you upload unreleased packaging or confidential assets.

People in your inputs. Section 4.2 says you may not submit "a photograph of another person without that person's permission." Get written consent for any real face you use as a character input.

Watermark. "All videos generated on a Free plan feature a Runway watermark" (Free plan details). Paid plans list "No watermarks".

Pricing, as published (USD, yearly toggle).

Source: Runway pricing. On the API, "Credits can be purchased for $0.01 per credit", with Gen-4.5 at 12 credits per second and Aleph 2.0 at 28 credits per second (API pricing).

Platform labels. Meta labels ad images "created or edited using third-party AI tools with AI info" (Meta). TikTok requires the AIGC label or your own clear disclaimer on AI-generated ad content, and counts "Showing the primary subject doing something they didn't actually do" as significant modification (TikTok ads policy). Aleph edits to a real person's footage fall in that group, so label them. More detail: AI-generated content labels on Meta and TikTok ads.

Runway in a one-script, every-format workflow

Runway fits the one-script method as the picture and re-versioning tool. Keep the script and voice track fixed, build a product demo with Gen-4.5, a mascot version with Act-Two, and test variants with Aleph 2.0 edits, then use a talking-head tool for the presenter read. The method is in one winning script, made in every ad format.

Anglr pays AI creators, editors and agencies on performance to remake winning ad scripts in every format: UGC, podcast, drama, animation and brand film. Anglr provides the winning scripts and angles, and every video gets a QC and compliance review before it goes live. If you already make full ads in Runway and know they perform, see how it works for creators.