Fundamentals
What is Stable Diffusion?
Stable Diffusion is an open text-to-image model that generates pictures by starting from random noise and repeatedly removing it, guided by your prompt.
Stable Diffusion is a latent diffusion model released publicly in 2022. The important word is public: unlike Midjourney or DALL·E, the weights were made available for anyone to download, run, modify and build on. That decision is why there are now thousands of fine-tunes, why it runs on consumer graphics cards, and why most free image generators on the web, Dreamify included, are built on top of it.
The way it works is less mysterious than it sounds. The model was trained by taking real images, adding noise to them in steps until they became static, and learning to reverse that process. To generate something new it starts from pure random noise and runs the reversal, removing a little noise at a time. Your prompt steers each step, nudging the result toward something that matches the description.
The word latent matters for speed. Rather than working on the full-size image, the model does all this denoising in a compressed representation, then decodes the finished result up to pixels at the end. That compression is what makes it possible to generate a good image on ordinary hardware in seconds rather than minutes.
Practically, being open means you get variety and control. Different fine-tunes of the same base model produce completely different aesthetics, and parameters like the seed, the scheduler and the negative prompt are exposed to you rather than hidden behind a conversational interface. It also means the model is less opinionated than a commercial product, which is why prompts for it tend to be longer and more specific.
Worth remembering
- It generates by removing noise step by step, guided by your prompt
- It is open, which is why so many free tools and fine-tunes exist
- It works in a compressed latent space, which is what makes it fast
- It is less opinionated than Midjourney, so prompts need more direction
Try it in the generator
The advanced panel exposes the negative prompt, scheduler, inference steps and seed. Twenty images free, no account needed.
Start generatingFrequently asked questions
- Is Stable Diffusion free?
- The model itself is openly available, so running it locally costs only your own hardware and electricity. Hosted services charge for the GPU time. Dreamify is free to try without an account.
- Do I need a powerful computer?
- Not to use Dreamify, which runs the models on cloud GPUs. To run Stable Diffusion locally you generally want a graphics card with at least 8GB of VRAM.
- How is it different from DALL·E?
- DALL·E is a closed model from OpenAI with stronger prompt adherence and much better text rendering. Stable Diffusion is open, more customisable, and exposes low level parameters that DALL·E hides.
Related terms
What is a diffusion model?
A diffusion model generates data by learning to reverse a noising process, turning random static into an image one denoising step at a time.
What is latent space?
Latent space is a compressed numerical representation of an image, where diffusion models do all their work before decoding back to pixels.
What is a LoRA?
A LoRA is a small add-on file that adjusts a model's behaviour toward a specific style, character or concept without retraining the whole model.
Back to the glossary, or apply it on a use case or an art style.