6 min read
How to Plan and Generate a Whole AI Photo Feed
One good photo is a start; a feed is what actually builds an audience. The canvas has a planning layer designed for exactly this: describe the kind of content you want, get back a structured shot plan you can cherry-pick from, and batch-generate the shots you choose in one go. Combined with chips that hold your character and look steady across every shot, a coherent week of posts takes minutes, not evenings.
step by step
Ask for a plan
Type /plan or just ask — "plan a week of coffee-shop lifestyle content for maya". The agent returns a plan card: a list of proposed shots, each with its own scene description.
Pick your shots
Select the shots worth making and skip the rest. Your plan selection limit scales with your tier, and you can always ask for a revised plan — "more night scenes, less latte art" — before committing.
Generate the batch
Hit generate and the selected shots render as a batch, landing on the canvas together (batch size scales by tier, up to 10 shots at once on Studio). Your chips apply to every shot, so the character, look and location logic stay consistent across the set.
Curate and finish
Press G to arrange your selection into a grid and review it like a contact sheet. Edit the near-misses, regenerate the duds, animate one or two winners into video, and download the set.
Plans beat one-at-a-time prompting
Prompting shot-by-shot produces a feed that wanders: every image is a fresh idea with fresh framing. A plan front-loads the creative direction — variety of scenes, consistency of story — and then executes mechanically. It's the difference between a photographer with a shot list and one wandering the city hoping. The plan card is also editable input, not a commitment: treat the agent's proposal as a first draft and shape it in conversation.
Drafting prompts with the agent
For hero shots that deserve more care than a batch, two helpers close the gap between idea and prompt.
/prompt — co-write mode
The agent drafts a full prompt you can edit line-by-line before approving generation. Nothing renders until you say so.
/describe — reverse-engineer
Point it at any image on the canvas and it writes a reusable shot prompt from what it sees — free of charge. Run it on your best shot, then vary the scene for the rest of the series.
Enhance mode
On by default, it upgrades every short message into a full prompt. Flip it off when you've hand-crafted exact wording and want it sent untouched.
Organizing a content pipeline
Creators shipping regularly tend to converge on the same structure: one canvas per series or per week, chips set once at the top of the session, a plan for volume, and hand-prompting only for hero shots. Finished sets get arranged into a grid for review; keepers are downloaded or animated. Because canvases persist, next week starts by duplicating the recipe — same chips, new plan.
quick tips
- Set your chips — character, outfit, location, shot settings — before generating the batch, so every shot in the set inherits them.
- Ask for more shots than you need and curate down; selection is cheaper than regeneration.
- Mix formats in one plan: a couple of portraits, a wide scene-setter, a detail shot — feeds read better with rhythm.
- Animate one or two batch winners into short clips for a mixed photo-and-video feed from the same session.
questions, answered
How many images can I generate at once?+
Batch size scales with your plan — from single shots on free up to 10 per batch on Studio. There's no limit on how many batches you run.
Can I change a plan after the agent proposes it?+
Yes — the plan is a conversation. Ask for different scenes, moods or quantities and it revises; you only generate the shots you select.
Will all the batch shots look consistent with each other?+
That's the point of chips: character, outfit, location and shot settings apply to the whole batch, so the set shares one identity and one look.
Do plans cost credits?+
Planning and prompt-drafting are conversation — credits are spent when images actually generate. Curating hard at the plan stage is free.
keep learning
building scenes →
Chips are draggable tiles — characters, outfits, locations, props and poses — that compose your next shot. Learn how to mix, link and toggle them to build 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.
shortcuts & workflow →
Every canvas keyboard shortcut and slash command in one place: pan, zoom, multi-select, grid arrange, and the / commands that drive images, videos, edits and plans.