Cladwell

Save an original retailer photo only when the person explicitly requests keeping the original. Uses an observed public image_url and product_url plus required clothing tags; transfers that exact source unchanged and does not create a flat lay. Default additions must prepare the image with ChatGPT, then use add_clothing_item with the actual finished file, even for already-clean retailer flat lays. Unavailable editing does not authorize this bypass. Establish the exact product/color match; do not invent links or guess variants. Private/account/order links are excluded. Saves directly to the selected/default capsule with the same free allowance and retry protection. A product page URL alone is not a photo.

add_clothing_from_product

Add one garment the person explicitly asked to add. Follow get_import_context image_preparation guidance first: ChatGPT prepares every new source as a faithful garment-only flat lay on white, unless the person explicitly requests the original. Requires the actual full-resolution finished host image and observed_details mapped to context tags. Use representative_illustration for generative reconstructions and omit the original crop on a prepared flat lay. Saves directly once complete, without an extra review or confirmation. Use top-level capsule_id/name, not the nested draft destination. Reuse the idempotency key and file identity for a retry; only saved=true confirms success. Missing required tags are returned without reserving allowance. Does not fetch personal closet context.

add_clothing_item

Read current connection and capability availability, without closet or style data. Item additions use the native allowance regardless of personal-context entitlement.

get_account_context

Read only the status and receipt of this person's direct clothing addition. Does not return closet contents, review drafts, private previews or any other person's operation.

get_clothing_addition

Get clothing tag choices, remaining import allowance, minimal destination identity/name and image_preparation guidance for an item addition. Call before saving: prepare every new image with ChatGPT's own image editor and the returned flat_lay_prompt, including clean retailer flat lays, then hand off the full finished file. Show the finished image inline and save without asking original versus prepared. Only an independently stated keep-original request bypasses preparation. Accept a named capsule or returned capsule_id. Never returns closet contents or style/history data. A name ambiguity returns choices without saving.

get_import_context

After the person selects the saved body photo and exact owned clothes for a requested visualization, return the Cladwell try-on prompt and ordered image references for ChatGPT's own image-generation capability. Accept the named/selected capsule and preserve exact garment identity/layering. This read does not generate an image, spend Cladwell inference credits, queue a job, or save a result. If host generation or reference transfer is unavailable, explain that limitation; never fall back to Cladwell's generation API.

get_try_on_context

When the person explicitly requests an outfit visualization using their saved photo, show their saved body-photo choices in ordinary chat. Requires current entitlement and separately approved photo sharing in chat. Do not choose a photo silently or load body photos for general advice. No inference, upload, save or provider call.

get_try_on_photos

Read account-wide Cladwell logged outfits, independent of the capsule selected for styling. Default past 365 days; use explicit start_date/end_date for older periods or all_history for all retained logs. Dates are inclusive UTC dates. Return up to 50 records per page; follow next_cursor with unchanged filters and limit for every remaining record. wear_counts_in_window covers the complete matching window; wear_counts_in_returned_outfits covers only this page. Favorites and previews are not wear logs. Missing logs never prove something was not worn. Read-only; no personal photos, location, notes, charges or saves.

get_wear_history

Read a bounded page of native outfit likes, Favorites, exact outfit exclusions, Outfits page star ratings with their reasons, or hidden outfits the person never wants suggested again. Requires current personal-context entitlement and separate preference read permission. Each source is distinct: no inference of dislikes from dismissal or missing wear. Preserve recorded capsule/date restrictions on exclusions; never suggest a hidden outfit's core combination. Page with unchanged kind/limit; no notes, chats, body photos or writes.

get_preference_context

Read saved style/color facts and up to eight owned item_ids in a named or returned capsule. Use capsule_name or capsule_id; otherwise the most recently used/main closet is selected. Carry the returned capsule reference in follow-ups. Ambiguous names return choices. Native capsule selection stays unchanged. Includes separately authorized account-wide wear history and preference evidence; follow their cursors for more. Missing saved results do not prevent advice. ChatGPT handles styling using these facts and the person's current preferences. Returns the saved photos for selected item_ids as inline image content; use this to show or inspect specific clothes. Report unavailable photos honestly. No review, quiz, outfit generation or saving.

get_styling_context