Stable Diffusion Client

This tool allows users to enter prompts and generates images using the backend self-hosted server of Stable Diffusion. With this tool, you can easily experiment with Stable Diffusion and create amazing images with just a few commands.

Setup

Step 1 - Host a Stable Diffusion Server on Banana Dev

Note: If you're a noob, I have already done that part. You just need to email me to get the API Keys.

Step 2 - Setup the client

Assuming you have python and pipenv installed. (Please see this if you want help with that).

pipenv install

Then create a .env file in the root folder with the API keys like so:

API_KEY=<your api key here>
MODEL_KEY=<you model key here>

Run

I have built a commandline tool to take in your prompt and give the output image

./draw "<Insert Your Prompt Here>"

Config

You can configure the size, negative prompts and other aspects of the image through config.json

Example config.json

{
        "endpoint": "txt2img",
        "params": {
            "prompt": argv,
            "negative_prompt": "cartoonish, low quality",
            "steps": 25,
            "sampler_name": "Euler a",
            "cfg_scale": 7.5,
            "seed": 42,
            "batch_size": 1,
            "n_iter": 1,
            "width": 512,
            "height": 512,
            "tiling": "false"
        }
    }

Dev Notes

Currently this tool has only been tested with the Automatic1111-stable-diffusion mod. The image saving feature might not work with other models since they give the outputs in different formats. I'm working on writing the adapters for those.

Roadmap

This tool is just made for testing and seeing whethdr there is a demand for such software in the market yet. Based on the results, I will work on this further if need be.

esxr/stable-diffusion-commandline-tool

Stable Diffusion Client

Setup

Run

Config

Dev Notes

Roadmap