AI Peeps has four image engines. You can pick one per image, and a Peep's appearance sets a default.
🔥 Forge — the flexible all-rounder
Forge is an image edit model: it can create images from scratch and edit existing ones, in any style — realistic, anime, painting, surreal, whatever. It understands natural language, so you just describe what you want.
Example prompt:
A woman in a red sundress sitting on a sunlit balcony, sipping coffee, warm morning light- Works with or without a Peep.
- The only engine that edits and combines images — it can blend up to 3 on-site images (Peep or generated) into one. See Editing & Combining.
📷 Photoreal — pure photorealism
Photoreal is our most lifelike engine. Like Forge it understands natural language, but it's tuned for maximum photorealism (real skin, light and detail). It doesn't edit, combine or make group photos (that's Forge's job), but it shines at single-character shots in any setting, from glam to fantasy.
Example prompt:
A woman in a red sundress sitting on a sunlit balcony, sipping coffee, warm morning light- Works with a Peep, or in Full Control mode with no Peep (just your own prompt).
- Works like magic as the first frame of a video for lifelike motion.
- You can set a Photoreal image as a Peep's reference model (see Peep from image).
✨ Prism — realistic & glam
Prism is realistic but highly stylized (not photorealistic): super-saturated colors and glamorous, eye-popping results. Instead of sentences it's driven by Booru tags, short comma-separated keywords like 1girl, red sundress, sitting. That style and vocabulary come from anime/art image boards called "boorus" (for example Danbooru and Gelbooru), whose community-tagged images trained these models, so their exact tag names are what Prism understands best.
Example prompt:
1girl, red sundress, balcony, morning light, sitting, coffee, looking at viewer- Browse a booru to find the exact tag for a pose, outfit or look, then paste the keywords in.
- You can weight a tag to emphasise it, e.g.
(red sundress:1.3). See the Prompting Guide. - Prism has the widest range of NSFW, which makes it a great pick for spicy work.
🧚 Anime — illustrated
Anime renders illustrated / anime art, also via Booru tags, with 14 sub-styles: Default, Vintage, Drawbook, Sketch, Retro, Detailed, Detailed Monochrome, Watercolors, Rainbow, Stilized, Pixel Art, Cute, Bimbo, and More Realistic.
Which engine should I use?
It comes down to the look you want:
- 📷 Photoreal: the most lifelike, real-photo results.
- 🔨 Forge: the flexible all-rounder, and the only one that edits, combines and makes group photos. Easiest to prompt.
- ✨ Prism: realistic but glam and super saturated, with the most varied NSFW (great for spicy work).
- 🧚 Anime: illustrated / anime art.
