Wan 2.7 Image Spicy Triptychs: Sequence Is Not Identity
A triptych needs three images with different jobs, not three near-duplicates placed side by side. Wan 2.7 Image Spicy exposes a sequential-output option for describing a group of images in one prompt. That can support a wide view, a portrait and a detail shot. It does not establish that the same face or garment will remain unchanged across them.
Key TakeawaysDescribe each intended image explicitly instead of asking vaguely for a coherent set.Siray's specification saysenable_sequentialturns on group output; with it enabled,nis the maximum image count.Count and inspect the returned images. A sequence request is not an identity lock, and assembling a finished triptych remains an editorial step.

What Does Sequential Output Mean Here?
As checked on October 15, 2026, the Siray Wan 2.7 Image t2i Spicy specification says to enable enable_sequential, state the intended number of images in the prompt and describe each image. In this mode, n sets the maximum number to generate; the prompt describes the intended group.
The specification also says billing follows images actually generated rather than that maximum. This is a count rule, not a price quote. Do not report n: 3 as proof that three usable images came back. Check the response and evaluate the outputs before calling the set complete.
Give Each Image a Different Job
Use the wide image to establish the setting, the portrait to show expression and the detail to emphasize a material or gesture. Keep the clothing and lighting descriptions consistent while changing framing. If every image changes pose, wardrobe, background and angle, the group has fewer shared visual anchors.
A coherent sequence can still contain an inconsistent person. Shared colors and a repeated room may make the images look related while concealing different facial features. Review identity separately from whether the three compositions tell a clear visual story.
Start with a non-explicit editorial subject and simple wardrobe. An illustrative group prompt:
Generate 3 separate editorial photographs of the same fictional adult aged 35, fully clothed in a dark jacket, in a quiet grey studio with soft light from the left. First image: wide seated view establishing the setting. Second image: shoulder-up portrait with a relaxed expression. Third image: close detail of the jacket sleeve and resting hand. Keep the intended hairstyle, jacket and lighting consistent across the group.
This describes a desired result, not tested model behavior. “Same adult” is an instruction to inspect, not a guarantee.
Keep the Request Small Enough to Review
For this exercise, use alibaba/wan-2.7-image-t2i-spicy, enable_sequential: true and n: 3. The required fields are model, prompt, size and aspect_ratio. A supported starting combination is size: "1k" with aspect_ratio: "3:4"; these are illustrative choices, not measured optimum settings.
n permits integers from 1 to 12. A larger maximum does not remove the need to specify the intended images. prompt_expansion_enable defaults to true; setting it to false removes the documented automatic prompt-expansion step. There is no negative_prompt in this image endpoint.
For the broader field layout, see the Wan 2.7 Image Spicy endpoint overview. Keep palette experiments separate from the first sequence test so a mismatch is easier to diagnose.
When Is the Triptych Ready to Assemble?
Check count, composition and continuity before making the final layout. Did the group actually include a wide view, a portrait and a detail? Are the hands plausible? Does the jacket retain its recognizable details? Do face and hair remain consistent enough for the intended use?
Reject an incompatible image rather than presenting it as proof of character consistency. Arrange accepted images in an editor with deliberate spacing; generated group output is not itself a promise of a finished three-panel layout.
Can a seed guarantee matching identity? No such guarantee is documented. A seed field does not replace reference review.
Is this reference-based editing? No. This t2i request does not upload a source portrait. Reference-image editing is a separate workflow.
Try a clearly described sequence through Siray, then decide which images genuinely belong together.
A Note on Compliance
All examples involve fictional or consenting adults aged 18 or older and remain non-explicit. Siray refuses illegal content and has zero tolerance for CSAM. Use reference materials only with required rights and consent. Siray provides access to third-party models; no self-hosting claim is made.
Source: Siray, Wan 2.7 Image t2i Spicy OpenAPI specification, retrieved October 15, 2026, linked above. No generation test was performed.