Models: create your AI double for photo posts
With a model, the Visual Designer can show you or your team in photo posts, no shoot needed. You need six required photos and your consent.
What is a model?
According to the app, a model is “an AI double of you or your team, built from a few targeted photos”. You create it in the Backstage under Models. In the Visual Designer it then appears in the Person field, but only for the Photo post type. If you choose a person there, they are fixed in the image. The app puts it like this: “the person IS in the image”.
No invented people
The Visual Designer doesn’t generate recognizable, invented faces. People in images come only from your model or from your own photo. Without a person, a photo image shows at most anonymous figures, for example from behind, blurred or as a silhouette.
How to create a model
Have these ready beforehand:
- An active plan or the trial. Without a plan, creating models is locked.
- Seven photos of the person, six of them required (see the table below).
- A phone or webcam if you want to take the photos directly in the browser.
- The consent of the person shown.
- Open Models. Choose Backstage › Models.
- Create a new model. Click New model and enter a name. The app asks “What should your model be called?”; examples are “Lea Business” or “Team look”. Afterwards it reports: “On to the photo session!”
- Fill the seven slots. Take the photos in the browser or upload them.
- Read the check result. Every photo is checked. For “Shot accepted”, you keep it with Keep or replace it with Retake. For any other result, you retake the photo with Try again.
- Complete the required photos. When the app shows “All required photos in. You can save.”, the six required slots are filled.
- Create. Click Create model now.
- Give your consent. In the “Final step: consent” dialog, you confirm the required checkbox.
- Wait. The app shows “Your model is being created”. When it’s done, it says, for example, “Lea Business is ready!”.
The seven photo slots
The slots are sorted into the groups Face close-ups, Half body and Full body. Six are required, one is recommended. Each slot shows its own tips in the app on what to look for and what to avoid.
| Slot | Required? | What matters |
|---|---|---|
| Front · neutral | Required | Like a good profile photo: camera at eye level, daylight, both eyes open, the face fills about half the frame. No sunglasses, no hat, no backlight, no beauty filter. |
| Front · smiling with teeth | Required | A real smile with visible teeth. According to the app, the AI otherwise doesn’t know your teeth. |
| Three-quarter left | Required | Head turned about 30°, looking at the camera. |
| Profile · side | Required | Head turned 90° to the side, looking straight ahead. |
| Hips up · front | Required | Portrait orientation, both hands visible, no second person in the frame. |
| Full body · front | Required | Head to toe, don’t cut off the feet. The self-timer helps. |
| Three-quarter right · smiling | Recommended | Mirror image of Three-quarter left, with a slight smile. |
Take photos, upload them and have them checked
Take or upload
- Take a photo: You take the picture with the camera in the browser. A live silhouette helps with alignment, and Switch camera changes to another camera.
- Upload or drag and drop: Drag the photo straight onto a tile. If you drop it elsewhere, it lands in the next free slot.
- JPG, PNG, WebP and HEIC are allowed, 25 MB at most. Every edge needs at least 1,024 px.
- The app recommends: “One photo at a time — the first one is checked.”
This checklist helps before every photo:
- Daylight, no backlight
- Camera at eye level, both eyes open
- No sunglasses, no hat
- No beauty filter
- Nobody else in the frame
- The same person in all photos
- Photo straight from your phone, so every edge is at least 1,024 px
The quality check
Every photo is checked right after you take or upload it (“AI is checking the photo…”). The result names strengths and weaknesses and shows under “How to do it better” what you can change.
| Result | Meaning |
|---|---|
| Shot accepted (green) | The photo works. You keep it with Keep or replace it with Retake. |
| Could be better (yellow) | The photo isn’t good enough yet. Under “How to do it better” you see what to change. You can retake it with Try again or cancel. Only photos marked Shot accepted are accepted. |
| Critical issue (red) | The photo has a serious problem. Retake it with Try again. |
| Identity mismatch (red) | The person doesn’t match the model’s earlier photos. Retake it with Try again. |
companycore doesn’t accept photos that fail the check: “This photo failed the quality check …”. The app counts how many required photos are still missing until it reports “All required photos in. You can save.” If you close the session without saving, it asks with “Photos not saved yet”. You then choose Discard or Keep editing.
Consent and other people
After Create model now, the “Final step: consent” dialog opens. A required checkbox confirms that the photos are processed as biometric data for the personal AI model. Without this checkbox, no model is created. How companycore handles the data is described in the Privacy Policy.
Other people: ask first
Only create models of team members or other people with their explicit consent. More on this: Photos of employees and customers.
After creating: use, add, delete
Use the model in the Visual Designer
The status shows you where the model stands: In progress with the number of photos, Not ready yet if required photos are missing, and Ready when it’s done.
- Open the Visual Designer.
- Choose the Photo post type. People are only available in this type.
- Choose the Person in the Photo setup. Only finished models are offered.
- Optionally choose an Outfit. The field appears as soon as a person is chosen.
- Click Generate now.
The photos of a finished model are fixed. For new photos, you delete the model with Delete model and create it again. When you delete, the app asks first and removes all photos of that model.
Wardrobe: outfits for your model
The wardrobe is a tab of its own in the Backstage (Backstage › Wardrobe). In the Wardrobe section you add garments, exactly one photo per item, sorted into Top, Pants / skirt, Shoes, Jacket / coat and Accessory. Under Outfits you combine items, at most one per category.
An outfit is assembled into one photo (“Compositing …”). That counts as an image generation. You can see how many you have per month under Settings › Usage. Background is in Usage and quotas.
Labeling for images with a person
Images with a person can look like real photos. For such images the Visual Designer shows the AI disclosure section, and the label “AI-GENERATED” sits at the bottom left of the image by default. How to change position and opacity, and what applies if you switch it off, is covered in AI labeling in companycore. The legal context is in AI labeling requirement.
Frequently asked questions
Why don’t I see a “Person” field in the Visual Designer?
The field only exists for the Photo post type. In addition, only finished models appear, and only if you have already created one in the Backstage.
Can I swap a model’s photos later?
No. After creation, the photos are fixed. Delete the model and create it again with new photos.
What does “Photo too small” mean?
Every edge of the photo must be at least 1,024 px long. Use a larger photo, for example straight from your phone.
The camera won’t start. What now?
Allow the camera in your browser or upload a photo instead. The app says: “No camera access. Allow the camera in your browser or upload a photo instead.”
What does “Identity mismatch” mean?
The person in the photo doesn’t match the model’s earlier photos. Check that the same person appears in all photos.
Was this article helpful?