Documentation Index
Fetch the complete documentation index at: https://docs.imagine.art/llms.txt
Use this file to discover all available pages before exploring further.
Summary
The Generate Video Node is the primary tool for creating videos using AI models in Workflows. It supports both text-to-video and image-to-video generation, giving you the flexibility to produce video content from a written description, a reference image, or a combination of both. You can choose from a wide range of AI models, each with its own strengths in style, motion quality, and output fidelity, and fine-tune settings like duration, resolution, aspect ratio, and more. Each AI model has different available parameters and supported settings, so it’s worth experimenting with the options to find the combination that best fits your creative vision.How to Use
Choose Your Model
Select an AI video model from the Model dropdown. Each model offers different strengths—some excel at cinematic motion, others at realistic physics or fast iteration. See the Video Models page for a full comparison.
Set Your Input
Text-to-Video: Connect a Prompt or AI Copilot node and describe the scene you want (e.g., “A drone shot flying over a misty mountain range at sunrise”).Image-to-Video: Connect an image from an Import, Generate Image, or Edit Image node to use as the first frame or visual reference, and add a prompt to describe the desired motion.
Configure Settings
Adjust Duration, Resolution, Aspect Ratio, and other parameters to match your output requirements.
Multi-Shot (Kling models)
Multi-Shot lets you chain multiple independently authored shots inside a single Kling generation, producing one continuous video sequence without manual stitching.Multi-Shot is available on Kling models only. Look for the Multi-Shot toggle inside the node settings when a Kling model is selected.
Select a Kling model
In the Generate Video node, choose any supported Kling model from the model dropdown.
Enable Multi-Shot
Toggle Multi-Shot on in the node settings panel. Shot slots will appear — Shot 1, Shot 2, Shot 3, and so on.
Author each shot
Write a separate prompt for each shot. Every shot has its own creative direction and is independently wirable — you can connect different upstream nodes to different shot inputs.
Set per-shot duration
Each shot supports a duration between 3 and 9 seconds. Adjust each shot individually to control pacing across the sequence.
Set Start and End Frames (optional)
Use the shared Start Frame and Last Frame inputs to lock visual continuity at the beginning and end of the full sequence. These apply across all shots in the generation.
Elements in Workflows
Elements are a persistent reference library for characters, objects, or any subject you want to keep visually consistent across shots. Once created, an Element can be referenced in any workflow via@mention in your prompt.
See Elements for full setup instructions, including how to create an Element and manage your library.
Using an Element in a Generate Video node:
Type @ in any prompt field and pick the Element from your library. It appears as a chip in the prompt and tells the model to lock that subject’s appearance for the generation.
