runwayml_character_performance

Technical Specifications of runwayml-character-performance

What is runwayml-character-performance?

runwayml-character-performance is CometAPI’s platform identifier for Runway’s character performance generation workflow, commonly described by Runway as Act-Two. The model is built for animating a character from a human performance clip: you provide a driving video of someone acting, plus a character image or video, and the system transfers facial movement, expression, speech timing, and body motion into a generated character shot.

This model is designed for creator workflows that need fast character animation without a traditional motion-capture setup. Runway describes it as a way to bring characters to life with realistic motion, speech, and expression, while also supporting non-human characters and varied visual styles. When the target character is a still image, the system can also synthesize environmental motion to make the shot feel more natural.

On CometAPI, this capability is exposed through the Runway character performance endpoint, making it suitable for product teams, creative tools, prototyping pipelines, and automated media workflows that want programmable access to performance-driven character animation. (cometapi.com)

Main features of runwayml-character-performance

• Performance transfer: Maps a person’s acting performance from a source video onto a target character, including facial movement, expressions, and broader body motion.
• Flexible character input: Accepts either a character image or a character video, giving developers different tradeoffs between motion control and scene preservation.
• Gesture control for image-based characters: When using a character image, the model can transfer gestures and body poses from the performance clip, enabling more expressive animation.
• Scene-aware image animation: For still-image character inputs, the system automatically adds environmental or camera-like motion so the result does not appear static or artificially frozen.
• Video-preserving character workflow: When the character input is a video, the generated result retains the original subject, scene, and camera motion while applying facial performance transfer.
• Adjustable facial expressiveness: Supports expression intensity controls, allowing developers to balance stronger acting transfer against visual stability and consistency.
• Production-friendly output options: Supports multiple common aspect ratios and resolutions at 24 FPS, making it usable for landscape, portrait, square, and cinematic social/video formats.
• Short-form animation optimization: Best suited for short clips, with a 3-second minimum and support up to 30 seconds, which fits ads, social clips, dialogue snippets, previs, and prototype scenes.

How to access and integrate runwayml-character-performance

Step 1: Sign Up for API Key

Log in to CometAPI. If you are not a user yet, register first. Then open your personal center, create an API token, and copy your key for authentication in subsequent requests. CometAPI exposes this model through its Runway-compatible API infrastructure. (cometapi.com)

Step 2: Send Requests to runwayml-character-performance API

Use CometAPI’s Runway character performance endpoint and set the model field to runwayml-character-performance in your application layer. The underlying request flow uses the character performance route with JSON input for the driving performance, character asset, motion controls, and output settings. (cometapi.com)

curl --location --request POST 'https://api.cometapi.com/runwayml/v1/character_performance' \
  --header 'X-Runway-Version: 2024-11-06' \
  --header 'Authorization: Bearer YOUR_COMETAPI_API_KEY' \
  --header 'Content-Type: application/json' \
  --data-raw '{
    "character": {
      "type": "video",
      "uri": "https://your-file-host.example/character.mp4"
    },
    "reference": {
      "type": "video",
      "uri": "https://your-file-host.example/performance.mp4"
    },
    "bodyControl": true,
    "expressionIntensity": 3,
    "seed": 4294967295,
    "model": "runwayml-character-performance",
    "ratio": "1280:720",
    "contentModeration": {
      "publicFigureThreshold": "auto"
    }
  }'

Step 3: Retrieve and Verify Results

After submission, process the API response in your backend or client application and verify that the returned generation matches your expected duration, framing, and expressive quality. In practice, you should also validate that the input assets follow Runway’s best practices—clear single-subject performance footage, visible face, stable framing, and suitable character references—to improve output reliability. (cometapi.com)