6 min read
How to Build AI Photo Scenes with Outfits, Locations & Props
Chips are the canvas's answer to prompt engineering. Instead of describing an outfit in forty words and hoping the model listens, you save the outfit once, drag its chip onto the board, and it's simply present in the next shot. Characters, outfits, locations, props and poses all work this way — and because they're tiles on a board rather than words in a prompt, you can see your whole scene at a glance and swap any ingredient in one click.
step by step
Build your asset library
Each panel in the left rail — Characters, Outfits, Locations, Props, Poses — lets you upload an image or generate an asset from a text description. Saved assets live in your library and can be reused on any canvas.
Drag chips onto the board
Drag an asset from the rail onto the canvas to make it a chip. Every active chip joins the next generation: character supplies the face, outfit the clothes, location the setting, props the objects, pose the body language.
Link outfits and props to a character
With multiple people in a scene, link an outfit or prop chip to a specific character chip so the assignment is explicit — the red dress goes on maya, the guitar goes to alex.
Describe the moment, not the ingredients
With chips in place, your chat message only needs the action and mood: "laughing mid-stride, paparazzi style". The agent composes chips plus description into the finished shot.
Why chips beat long prompts
A prompt is ephemeral — you write it, it generates, it's gone. A chip is persistent state. Set your scene once and every subsequent message inherits it, which is what makes shooting a series practical: ten shots in the same location with the same outfit is ten short messages, not ten repeated paragraphs. It also makes changes surgical. Swap the location chip and regenerate; nothing else moves.
The two icon chips
Beyond asset chips there are two settings chips, added from the rail. The shot chip holds photographic settings — framing, lens, lighting, color grade. The camera chip holds motion settings for video — moves, rigs, speeds. They render as icon tiles with an edit gear, and like all chips they can be toggled or removed per shot. They get their own tutorials; here it's enough to know they compose with everything else.
Toggle instead of delete
Every chip has an active toggle. Deactivate the location for one indoor shot, then flip it back on — no need to re-drag anything.
Poses direct body language
A pose chip pins posture and framing — seated, walking, over-the-shoulder — which is much more reliable than describing body position in words.
Props stay in hand
Save recurring objects — a coffee cup, a handbag, a skateboard — as props so they look identical across the whole series instead of being reinvented per shot.
The starter collection
You don't have to build a library from nothing. The starter collection offers ready-made characters, outfits, locations and props that subscribers can claim into their library and remix immediately. It's the fastest way to see how a fully-dressed scene behaves before you invest in crafting your own assets.
quick tips
- Name assets clearly when you save them — "black leather jacket" beats "outfit 3" when the agent references them in prompts.
- Keep a canvas per project: chips persist on the board, so a dedicated canvas keeps your café series from inheriting last week's beach location.
- Generate location assets from text for places you can't photograph — "neon-lit Tokyo alley at night, rain" makes a reusable location chip.
- Save limits apply when saving new assets (free: 5 locations and props, 3 characters), never when generating — you can always shoot with what you have.
questions, answered
What's the difference between a chip and a saved asset?+
The asset lives in your library permanently; a chip is that asset placed on a specific canvas. Delete a chip and the asset is still in your library, ready to drag onto any board.
Do all chips affect every generation?+
Every active chip does. Use each chip's toggle to sit one out — an inactive chip stays on the board but is ignored until you flip it back on.
Can I generate outfits or locations instead of uploading photos?+
Yes — every asset panel supports generating from a text prompt as well as uploading an image. Generated assets are saved to your library like any other.
How exact is outfit matching?+
Very close, especially when the outfit chip stays linked to the same character across shots — the studio treats the dressed character as one subject, which keeps garment details stable.
keep learning
getting started →
Learn the selfielab canvas in five minutes: drop a character onto the infinite board, chat with the agent, and generate your first AI photo — same face, any scene.
shot settings →
Framing, lenses, lighting, color grades and camera styles: use selfielab's shot settings chip and presets like iPhone Candid or Studio Editorial to shoot like a photographer.
consistent characters →
Character consistency is the whole point of an AI photo studio. Here's how selfielab keeps one face across every shot — uploads, personas, styling, and reference images.