A Node.js CLI for generating AI images with Replicate and saving them to disk.

• Easy to install and use.
• Supports any model on Replicate that has a prompt input. Default is Flux Schnell.
• Saves all the files to disk for you.
• Includes prediction id and a slug of the prompt in the filename.
• Lets you specify an exact prompt, or specify a subject and roll the dice with random promptmaker prompts.

npm i -g aimg

Then grab a Replicate API token and set it in your environment:

export REPLICATE_API_TOKEN="r8_..."

All that is required is a prompt:

aimg "cute cat"

This will generate three images of a cute cat and save them to the current directory.

If your prompt is long, you can put it in a file:

aimg "$(cat prompt.md)"

The default model is Flux Schnell, but you can specify any model with the --model option.

aimg "cute cat" --model bytedance/sdxl-lightning-4step

This will work for any model on Replicate that takes a prompt as input and outputs a list of URLs.

Examples:

• black-forest-labs/flux-dev
• bytedance/sdxl-lightning-4step
• stability-ai/stable-diffusion-3
• zeke/ziki-flux ← works for Flux fine-tunes too!
• more...

The latest version of the model is used unless you specify a version in the format {owner}/{name}:{version}:

aimg "cute cat" --model zeke/ziki-flux:dadc276a9062240e68f110ca06521752f334777a94f031feb0ae78ae3edca58e

If you want more images, use the --count option:

aimg "cute cat" --count 20

If you want generate a different semi-random prompt for each image, specify a subject and it will use promptmaker to generate random prompts:

aimg --subject "a white cat"

If you specify a subject, then prompt is ignored.

If you want to save the images to a specific directory, use the --outputdir option:

aimg "pink cat" --outputdir "pink-cat"

Hee's an example that loads a prompt from a file, generates 100 images with random "pink cat" prompts, and stuffs them in a specific directory:

aimg --subject "pink cat" --count 100 --outputdir "pink-cat"

• --count: Number of images to generate. Defaults to 3
• --model: Model to use. Defaults to stability-ai/stable-diffusion-3
• --subject: Subject to pass into promptmaker to generate random prompts