Omost: First Look

Omost: First Look


Omost is a cutting-edge AI tool that transforms the coding capabilities of Large Language Models (LLMs) into powerful image generation and composition capabilities. It works seamlessly with RunDiffusion’s cloud platform, Omost offers a versatile and efficient way to create high-quality visual content. In this guide, we’ll explore the features of Omost, how it works, and how you can use it to enhance your image creation projects.

What is Omost?

Omost is an AI project that leverages the coding capabilities of LLMs to generate and compose images. The name "Omost" signifies two things: it means "almost" because your image is almost there after using the tool, and it represents "omni" (multi-modal) and "most" as it aims to get the most out of its capabilities.

Omost utilizes pretrained LLM models that write code to compose visual content on a virtual Canvas agent. This Canvas is then rendered by specific image generation implementations to produce the final images. The models are trained using a combination of ground-truth annotations, automatically annotated data, reinforcement learning, and tuning data from OpenAI’s GPT-4.

Key Features of Omost

  1. Versatile Image Composition: Omost allows for complex image compositions using the coding capabilities of LLMs.
  2. Pretrained Models: It comes with three pretrained models based on variations of Llama3 and Phi3, ensuring robust performance.
  3. Virtual Canvas Agent: The virtual Canvas acts as a workspace where visual elements are composed before being rendered into final images.
  4. Multi-Modal Capability: By integrating multi-modal data, Omost can generate diverse and high-quality visual content.

Getting Started with Omost on RunDiffusion

Step 1: Accessing Omost

To start using Omost, log in to your RunDiffusion account. Navigate to the tools section and select Omost from the list of available AI tools. You will want a Large server.  You won’t need to change the settings when you start so just scroll down to the text box once it launches.

Step 2: Entering your first prompt.

My prompt: A portrait of a beautiful woman with freckles and braids wearing glasses from Japan.  I didn’t change any settings because I love the Juggernaut X model and knew it would work just fine.  I wanted to try it on default before I got fancy so I left the steps and everything as is.

As you can see it did as asked and created a nice prompt on my behalf kind of super charging it and the final prompt aligned well with the Juggernaut X model and created that beautiful image. Now here is the part where Omost shines in my opinion.  You can have a conversation with the LLM to create your desired end result. 

You can see it has kept the freckles from my first prompt. I lost the nose ring though.

I tested it with very natural language. If you follow their github they give recommendations on the best way of communicating with the models to take it to the next step.  But I enjoys peaking with LLMs in a very natural way.  And the results were beautiful.  Tips for Optimal Use

  • Start with Simple Compositions: Begin with basic compositions to understand how Omost’s coding and rendering process works.
  • Save Your Work Regularly: Regularly save your progress to avoid losing any changes during the composition process.


Omost is a versatile and powerful tool that brings a new dimension to image creation by combining the coding capabilities of LLMs with advanced image generation techniques. By integrating Omost with RunDiffusion’s cloud services, you can leverage its full potential without the need for local resources. Whether you are a professional artist or a hobbyist, Omost provides an accessible and efficient way to create stunning visuals. Start exploring Omost today and unlock new creative possibilities on RunDiffusion.

For further reading and tutorials on related tools, check out our other articles:

About the author
Adam Stewart

Great! You’ve successfully signed up.

Welcome back! You've successfully signed in.

You've successfully subscribed to .

Success! Check your email for magic link to sign-in.

Success! Your billing info has been updated.

Your billing was not updated.