Overview
Runway is a cloud-based platform for generating and editing video, images, and audio with AI. Its core web app lets users generate video from text prompts, still images, or existing footage, then refine results with a set of purpose-built editing tools rather than a traditional timeline editor alone. Runway also runs a separate developer platform with a public API, so its models can be called programmatically and combined into custom pipelines outside the consumer app.
AI capabilities
Runway's flagship video model line is Gen-4.5, described on its own site as its highest-fidelity model for text-to-video and image-to-video generation, with support for additional guided inputs. Alongside it, Aleph is positioned as an in-context video editing model: used for tasks like restyling a shot, changing lighting, or adding and removing elements from existing footage rather than generating a clip from scratch. A separate model, referred to as Act-Two, is used for performance- and motion-driven generation. Runway's site also describes a general world model, referred to as GWM-1, aimed at simulation and robotics use cases rather than creative content. The consumer app also bundles task-specific tools such as background removal, video upscaling, dialogue/audio addition, and scene relighting, alongside a node-based workflow builder (Workflows) and a conversational "Agent" mode for producing a video end-to-end from a described brief. Image generation and editing are also available in the same workspace.
Editing tools and workflow
A distinguishing feature of Runway relative to many single-purpose generators is its emphasis on editing existing video rather than only generating new clips. Its editing tools operate at something closer to a frame or shot level (for example, applying a style change to one frame and propagating it across a clip), which is a different workflow than the keyframe-driven or frame-interpolation approaches used by some other video generators. Runway also documents motion control and temporal consistency as explicit design goals of its editing tools, terms that have become common vocabulary across the AI video space more broadly.
Access model and API
Runway is accessed through a subscription-based web app, with generation limits and higher-tier features scaling by plan. Separately, Runway operates a developer API (Runway Dev) through its own developer portal, billed on a credit-based, usage system that is independent from the consumer app's credit pool. A developer can create an account, add credits, and start making API calls directly, without an approval or waitlist process; per-generation cost varies by model, resolution, and clip length, since longer or higher-fidelity generations consume more credits than short, low-resolution ones.
Content policy and adult-content positioning
Runway is a mainstream creative tool, not an adult-content platform. Its published usage policy explicitly prohibits sexually explicit content, adult nudity, and non-consensual intimate imagery in both inputs and outputs, and states that violations are enforced through a combination of automated detection and human review. It also prioritizes and reports any child sexual abuse material it becomes aware of. Runway's own Characters product additionally bars building characters based on the likeness of minors and bars content designed to engage users under 18.
Given this policy, Runway itself has no role as a destination for adult content on this site. It's documented here because its models, editing paradigm, and the general vocabulary of AI video generation it has helped popularize are referenced across the wider generative-video ecosystem, including by tools and communities that do work with adult content using different, less restrictive models.
Limitations and important distinctions
Because Runway enforces a no-nudity, no-sexually-explicit-content policy at the platform level, it cannot be used for adult image or video generation regardless of subscription tier or API access; this differs from Civitai or open-weight tools built on models like those hosted on Hugging Face, where adult content is possible because the underlying weights and hosting are not centrally policy-gated in the same way. Runway is also entirely hosted; there is no local/self-hosted deployment option comparable to running an open-weight diffusion checkpoint on local hardware.
Technical context
Runway’s unpublished internals are not inferred; this dossier explains only the documented video design space.