The open source picks are flagged. Those are the ones you can run yourself for free, with no monthly fee. I also mark anything that has shut down, so you do not chase a dead link.
New to running your own? Start with these:
The open model that started it all. Free to run yourself, endlessly customizable, and the base for half the tools on this page.
Visit site →
The image quality is still the best I have seen. Paid only, runs on Discord and on the web.
Visit site →
OpenAI's current image model, built into ChatGPT. Strong at following a long, detailed prompt.
Visit site →
Adobe's generator, built into Photoshop as Generative Fill. Trained on licensed images, so it is safer for commercial work.
Visit site →
A friendly all-rounder with its own models, presets, and a generous free daily allowance.
Visit site →
The one that actually gets text right inside an image. Great for logos, posters, and memes.
Visit site →
Beginner friendly with a real free tier. Quality is a notch below the top tools.
Visit site →
A friendly hub with several models, plenty of styles, and a big community. Runs on credits.
Visit site →
Stability AI's web app for Stable Diffusion, the open model. You buy credits and get a lot of control.
Visit site →
A hub for many Stable Diffusion models, ready made workflows, and even training on your own images.
Visit site →
Blend and 'breed' images to morph faces and characters. Great for exploring variations.
Visit site →
A social gallery for AI art, with prompt sharing and creator tools. Leans adult.
Visit site →
Free and fast, quality is rough. Good for a quick laugh or a rough idea.
Visit site →
Generate random human faces with the traits you choose. By Generated.photos, from a licensed dataset.
Visit site →
A cheap utility belt: AI headshots, background removal, photo restoration, and quick templates.
Visit site →
A small Japanese tool for turning photos into manga and cleaning up backgrounds.
Visit site →
The image side of Jasper's marketing suite. Fine for quick commercial visuals if you already use Jasper.
Visit site →
A Japanese illustration tool that learns a character from your own reference art.
Visit site →
One click and you get a fresh AI face. Nothing to set up.
The classic demo. Every refresh gives you a brand new face that is not real.
Visit site →
OpenAI's older image model. These days DALL-E 3 inside ChatGPT is the one to use.
Visit site →