Actions16
- LLM (Text Generation) Actions
- Image Generation Actions
- Video Generation Actions
- Text-to-Speech Actions
- Speech-to-Text Actions
- Music Generation Actions
- Embeddings Actions
- Content Moderation Actions
- Custom Inference Actions
Video Generation → Keyframe Interpolation
AI-generatedSummary
Interpolate a video by generating frames between multiple keyframe images based on a text prompt, producing a smooth animated video sequence.
Inputs
- Chute — Select a specific chute to use or enter a custom chute URL; determines the video generation model.
- Prompt (required) — Text description of the desired video or animation, guiding the interpolation style.
- Keyframe Images (required) — Multiple keyframe images each with an image URL or binary data, frame index (position), and influence strength (0.0 to 1.0) specifying their effect on the generated animation.
- Additional Options — Optional parameters including video resolution, number of denoising steps (quality vs speed), video duration (seconds), frames per second (fps), guidance scale (adherence to prompt), random seed for reproducibility, negative prompts (what to avoid), pipeline type (should be 'keyframe_interp' for this operation), LoRA adapters for style transfer, and a maximum timeout for the generation process.
Output shape
a single binary video file or JSON containing the generated video data or URL, depending on the selected chute’s response format.
Output format depends on chute capabilities and response; typically includes raw video binary data with metadata or a JSON object with video URL. The node waits synchronously with an optional timeout. At least two keyframe images are required to interpolate a video successfully.