ChatGPT AI Guide: Write a Dual-Reference Kurta-Pyjama Specification

By TryBuy Editorial

Use ChatGPT to turn separate kurta and pyjama records into a dual-reference image specification. The output should bind each visible feature to the correct garment and preserve the relationship between the pieces. This guide prepares text for a production workflow; it does not claim that ChatGPT has generated, edited or physically tested the complete outfit.

Supply two garment records and one relationship note

Label the top TOP-A and the bottom BOTTOM-B, or use your project’s actual identifiers. Include the source descriptions, available view labels and documented visible features for each. Keep the waistband, leg shape and material information for the bottom separate from the kurta’s collar, sleeves and decoration. Leave hidden fields unknown.

Add whether the pieces are a confirmed sold set, separately owned garments or a proposed styling combination. Copy inclusions from the exact current listing when they matter. A coordinated photograph is not a substitute for this relationship note. The model should not infer that the two garments share a fabric or that both arrive in one package.

Copy the specification prompt

Create a dual-reference image specification from TOP-A, BOTTOM-B and my relationship note. Return separate protected-feature lists, source pointers, view limitations, planned pose and crop, and cross-transfer errors to reject. Keep every feature attached to its original garment. Do not infer hidden waistband construction, transfer motifs or material claims between pieces, change package contents or claim physical fit. Draft a bounded production brief and a review checklist; do not describe the image as completed or approved.

Ask for a forbidden-transfer section. This makes the specification more actionable than a general preserve everything instruction while keeping its content grounded in the actual records you supplied.

Filled fictional example

Imagine TOP-A has a documented decorative pattern and BOTTOM-B is a separately owned plain garment. The fictional specification can state that the top’s pattern must not appear on the bottom. It should not name the bottom’s fabric if that information is missing or invent a waistband hidden under the kurta.

If the planned pose obscures the top’s lower border, the output should mark that region not assessable and ask whether another view is needed for the intended article. It cannot recover the concealed details through text. The example demonstrates a written specification, not a successful image result or a tested claim about reference-handling tools.

Revise the specification by garment

• “List every field that lost its TOP-A or BOTTOM-B source pointer.”
• “Remove any inferred shared fabric, closure or inclusion.”
• “Rewrite the production brief for the new crop while keeping protected features and relationship facts unchanged.”

If an actual candidate is later reviewed, supply human observations separately. Ask for a discrepancy log rather than letting the model invent the result that the specification intended to produce. A plan and an image review are different records.

Validate the handoff

Check both protected-feature lists against the exact sources. Confirm that the relationship note matches the listing or your owned-outfit record. Review whether the requested crop can show the feature central to the article. Keep account, model and editing controls out of the brief unless you have verified the production tool’s current capabilities for the task.

Ask a human editor to confirm that every source label points to the intended file before any image-production step begins.

Use the original practical guide for the practical two-reference workflow and the men’s kurtas collection for current configurations. Your finished deliverable is a source-bound text specification, a forbidden-transfer list and a review checklist. It should make the next production step clearer without presenting an AI-written brief as evidence that the two garments have been combined accurately or that any package inclusion has changed.

For general instructions and follow-up techniques, see official OpenAI prompting guidance.