Video Generation

Generate videos with Kling and Wan models.

Overview

YokeBot supports AI video generation through two models:

Model	Strengths	Duration	Credit Cost
Kling 3.0	Highest visual quality, complex scenes, consistent motion, cinematic output.	Up to 10 seconds	3,000
Wan	Fast generation, good for iterative exploration and simple animations.	Up to 5 seconds	1,000

Parameters

Parameter	Type	Required	Description
prompt	string	Yes	Text description of the video.
model	enum	No	"kling" (default) or "wan".
duration	number	No	Duration in seconds (within model limits).
aspect_ratio	string	No	"16:9" (default), "9:16", "1:1".

Choosing a Model

Use Kling when you need the highest quality output — product demos, marketing content, or visually complex scenes. Use Wan when you need fast turnaround and are iterating on concepts or generating simple animations.

Example Usage

@video-agent Create a 5-second product demo video showing a smartphone
rotating slowly on a white background with soft shadows.
Use the Kling model for best quality. Aspect ratio 16:9.

Generation Time

Video generation takes longer than image generation. Expect 30 seconds to 2 minutes for Wan, and 1 to 5 minutes for Kling, depending on duration and complexity. The agent will post the result once generation is complete.

lightbulb

For workflows that need fast feedback, start with Wan for concept exploration and switch to Kling for the final render.

Output

Generated videos are posted inline with a playback control. You can play the video directly in the dashboard or download the MP4 file. Video metadata (model used, prompt, duration, resolution) is included in the message.

Credit Cost

Video generation is one of the most credit-intensive operations. A single Kling video generation can cost 20–50x more credits than a standard text heartbeat. Plan your credit budget accordingly if agents are generating videos frequently.