How do I reuse the same AI fashion model across an entire collection?

Quick Answer

Nightjar carries one AI fashion model across a collection by saving the person once as a Fashion Model and selecting that same identity reference for every product, instead of describing the person again each time. You don't need to train a LoRA or fine-tune an image model: choose one of Nightjar's 80+ pre-built Fashion Models, or create a custom one from 1 to 5 photos of a real person you have permission to use. Save it in a Recipe with the Pose, crop, photographic look, and setting, then apply that Recipe to each Product. The Fashion Model supplies the person, each Product's own photos supply the garment, and the Recipe keeps the shoot direction fixed while the merchandise changes. This improves identity continuity but cannot promise an exact face or body across all images, so review every output.

How do I reuse one Fashion Model for every Product in Nightjar?

Nightjar's repeatable collection workflow separates who appears from what is being photographed. A Product groups the Product Photos and factual details for one visually distinct sellable item, while a Fashion Model supplies one reusable AI-person reference.

Create a Product for each distinct item or colorway. Add clear Product Photos that show important garment details and choose the best identity-defining image as its Main photo.
Open Generate > Create, select the first Product, and choose With model under Shot > Subject.
Open Model and select one of Nightjar's 80+ pre-built Fashion Models. To base a custom Fashion Model on an authorized real person, choose Create New and add 1 to 5 images of that same person.
Set the reusable visual direction. A Pose controls body arrangement, Camera Distance controls the crop, a Photography Style controls the photographic look, and a Background controls the setting.
Save the setup as a Recipe, Nightjar's reusable Create-form setup. The Recipe stores the Fashion Model and production direction, but not the Product or its photos, so you can select the next Product and reapply the same setup.

For a collection made by several people, Products, Fashion Models, and Recipes belong to the Team. Collaborators can therefore work from the same approved identity and setup instead of rebuilding the brief. For programmatic catalog production, the Nightjar API exposes the same Products, Fashion Models, and Product Photography Workflow, but every output still needs review.

To run the whole collection from one request instead of one Product at a time, connect Nightjar to Claude, Claude Code, or Codex through MCP, the open standard AI assistants use to connect to other tools. Ask, for example: "Apply our Linen Recipe to every product in the Summer collection. Show me three samples first." Nightjar doesn't know your store's collections, so name the Products or let the assistant match them from your saved list. The assistant plans a batch of up to 100 Products, each with its own Generation using the Recipe's Fashion Model and settings. Planning is free and shows the Credit cost and a spending ceiling before anything runs. Choose samples that cover the garment types you plan to shoot, compare them with the approved identity reference, and only then approve the rest. Running a catalog batch from an AI assistant walks through each step, and Making product photos in Claude with Nightjar shows the conversation.

Should I use Create, Edit, or Photoshoot to keep the same fashion model?

Use Create with a Product, Fashion Model, and Recipe for repeatable collection production. Edit and Photoshoot solve different jobs.

Nightjar path	Best use	Identity and collection behavior
Create: Single shots	Repeating a controlled shot across Products	Reuses the selected Fashion Model, Pose, Camera Distance, look, setting, and output choices through a Recipe.
Create: Photoshoot	Making one cohesive four-image set	Keeps the selected subject and direction connected, but deliberately varies pose, angle, crop, and detail. It is not the choice for an identical catalog pose.
Edit	Reworking or combining existing images, including a one-off Try On	Uses explicit image references and written direction. It does not apply Create's reusable Fashion Model, Pose, or Camera Distance controls.

If a Recipe references a Fashion Model that has since been deleted or is unavailable, Nightjar applies the remaining setup, omits that ingredient, and reports the omission. Confirm that the intended Fashion Model is present before starting the next Generation.

Do I need to train a LoRA or fine-tune a model to keep the same person?

No. In Nightjar, keeping the same person across a collection takes a saved Fashion Model, not a training run. A LoRA adds a small set of new weights to one image model and trains them on your images (Hugging Face's LoRA training guide), and a fine-tune trains the model itself. Either way, you prepare training images, run the training, and use the result with the model it was trained for. A custom Nightjar Fashion Model starts from 1 to 5 photos of the same person instead. Nightjar generates one identity card of the person's face and hair for 1 Credit, and every later Generation that selects that Fashion Model uses the card as its reference. Nightjar does not train or load LoRA files.

For a collection, the larger difference is everything else that has to stay the same. A LoRA of a person addresses only who appears; every image still has to show the real garment accurately and repeat the same pose, crop, light, and setting. Nightjar keeps those as separate saved inputs: the Fashion Model for the person, each Product's photos for the merchandise, and the Recipe for the shoot direction. That is what lets one setup carry from the first Product to the last: each new Product starts from the approved person and shoot direction rather than a new brief. If the person is your brand's real model, creating a Fashion Model from a real model explains what the identity card carries and which permissions to get first.

Does reusing a Fashion Model preserve an exact identity?

Reusing a Fashion Model means reusing one identity reference, not fixed facial pixels. A Fashion Model's identity card, pre-built or custom, shows only the face and hair, so build, tattoos, and hands are not saved with it; if they must match across the collection, write them into the Recipe's Custom Directions. Pose, crop, expression, lighting, hair, clothing, product contact, and generative variation can still change how the person appears. A custom Fashion Model can help an authorized real person remain recognizable, but it is not an exact digital replica or a fit-prediction avatar.

Nightjar's built-in visual review can retry an obvious eligible failure, such as a plainly substituted Fashion Model, without another Credit. That review is additional protection, not fine-grained likeness, anatomy, garment-fit, or product approval. Compare each output at full size with the approved identity reference and Product Photos, checking the face, body, hands, garment construction, logos, patterns, color, fit, and contact points.

What source and likeness rights do I need for a custom Fashion Model?

Use source photographs only when you have the right to upload and process them. If a custom Fashion Model is based on a real person, obtain permission that covers creating and commercially reusing the AI likeness in the intended media, markets, and time period; clear the source photographer's rights separately when needed. Nightjar's Terms of Service prohibit Input or Output that infringes third-party privacy or publicity rights, so access to the feature does not supply those permissions.