Start a ComfyUI diffusion workflow with a blank canvas in latent form. Add Empty Latent Image through Add Node > latent, then send its LATENT output to KSampler. section 8.2.4
Configure the initial image here:
- width: the generated image width section 8.2.4
- height: the generated image height section 8.2.4
- batch_size: how many images the workflow prepares together section 8.2.4
Basic connection
Create Empty Latent Image, set the desired width, height, and batch_size, and connect LATENT to KSampler’s latent_image input. section 8.2.4
Figure 8.12 The Empty Latent Images is where we specify image dimensions and batch size.

Use the figure to identify the node controls for image dimensions and batch size. section 8.2.4
