Connect a Runway model to an app with AI
Integrate a chosen media model through the official SDK. Handle the actual input fields, task progress and temporary output links in the application.
To connect a Runway model to an app with AI, tell runway-dev-models which model the team has selected and where the user submits the material. Two video models can have different inputs, supported options and billing units. A working call to a neighbouring endpoint is a poor substitute for reading the contract that applies to this one.
The skill guides the server-side integration and the states that the user sees while media is being produced. If generation fails, the application can show the failure without quietly submitting a second paid request. Signed output links are temporary, so the team needs a storage plan when a result should remain available after the session.
What it does
- Check that the chosen project can access the model.
- Inspect its actual inputs, outputs and constraints.
- Connect the official SDK on the server.
- Track a generation task and handle its failures.
- Plan storage for outputs needed beyond a temporary URL.
Install and use
- Install the skill in a compatible agent. A Runway Dev account, server-side API secret and official SDK are needed for a live integration; generation can use credits.
- Ask “Connect our chosen image model to this page. Check its reference-image options first and validate locally before a paid call.”
- Review loading and failure states and decide which results the application should retain.
FAQ
Source
Made by Runway. Original skill: runway-dev-models, published on GitHub under the MIT licence.
View on GitHub