Introduction
Wan 2.7 Text to image turns a written art direction into a new image. Use it when composition, subject, lighting, and visual hierarchy need to be designed from a brief.
Choose a Wan 2.7 mode by starting material
| What you have at the start | Mode |
|---|---|
| a written visual brief | Wan 2.7 Text to image |
| an image that needs a directed change | Wan 2.7 Image to image |
Key features & highlights
The 9-image edit draws what you wrote
Give image-to-image a reference and it draws the brief you sent, wardrobe and framing included. There is no content flag and no per-request declaration anywhere on this platform, on either endpoint — the difference is the model. On text-to-image it answers a restricted prompt with a more conservative frame instead. The cafe scene for this endpoint is text-to-image, an oversized white shirt with rain on the window, and that is roughly where the text endpoint stops.
Combine up to nine images without painting a mask
Image-to-image takes one to nine source images in a single request, so a garment, a face, a room and a colour reference can all enter the same edit. Say what each source is for in the prompt — an unlabelled stack gets averaged, which is the most common reason a nine-image edit comes back muddy.
It plans the layout before it draws
Wan 2.7 plans the frame before it draws it, which shows up in things that need a spatial decision: who is in the foreground, what the hand is holding, where the light source sits. Long descriptive prompts with a clear subject and one clear relationship beat keyword piles here.
Request parameters
| Parameter | Required | Description |
|---|---|---|
prompt | Yes | Describe the frame: subject, action, camera position, then light and lens. Wan 2.7 plans composition first, so one clear spatial relationship beats a pile of adjectives. at least 1 characters · at most 5000 characters |
aspect_ratio | No | 1:1, 4:3, 3:4, 16:9, 9:16, 3:2 or 2:3. Defaults to 1:1. |
enable_thinking | No | Lets the model reason about the prompt before it starts generating. Turning it off is noticeably faster and usually costs nothing in quality on detailed prompts. Default: |
num_outputs | No | 1 to 4, each a separate seed on the same brief. Four readings of one layout are directly comparable because nothing but the seed moved, and each image is billed. |
resolution | No | 1k or 2k; 1k by default. 2k carries four times the pixels of 1k, and the shape still follows |
color_palette | No | Optional colour theme of 3 to 10 colours (8 is the recommended count). at least 3 items · at most 10 items |
seed | No | 0 to 2147483647. Hold it and change one clause at a time: the same seed keeps the draw steady, so the clause is what moved. |
Pricing
$0.03 / image
Use cases
Explore a visual concept before production
Text-to-image is the endpoint for a frame with nothing fixed yet: no face, garment or room that has to survive. Write subject, action, camera position, light and lens, and ask for up to four readings at once with num_outputs; each is a separate seed on the same brief and is billed as its own image.
Create a storyboard key image
Generate the still at 3:2 or 9:16, then hand it to an image-to-video endpoint as the opening frame. The video prompt then only has to carry motion, which is the most literal way to get an exact wardrobe and pose into a clip.
Establish a campaign visual system
Hold one seed and one prompt skeleton, then sweep aspect_ratio across 1:1, 3:2 and 9:16 for a profile, a banner and a story card. The pixel budget is the same for every ratio, so a wider frame is wider, not larger.
Prompt guide & pro tips
Write composition before decoration
Name the subject, what they are doing and where the camera sits before light, materials and style. Wan 2.7 plans the frame before drawing it, so one clear spatial relationship beats a pile of adjectives.
Fix the seed, then change one thing
Once a frame is close, hold seed and vary a single clause: the wardrobe, the light, the ratio. The same seed keeps the draw steady, so any difference you see is the clause you changed; it does not guarantee an identical image.
Check the canvas before reusing the image
At the 1k default the canvas is about 1.01 to 1.05 megapixels whichever ratio you pick, and resolution: 2k gives about four times that. Text alone invents a new face on each seed, so when someone has to stay recognisable, move to image-to-image with a reference. Turning enable_thinking off is noticeably faster.
API quick start
# Generate the idempotency key once; reuse it when retrying after a lost response.
export SPICY_IDEMPOTENCY_KEY="${SPICY_IDEMPOTENCY_KEY:-$(uuidgen)}"
curl -X POST https://api.spicyapi.ai/api/v1/jobs/createTask \
-H "Authorization: Bearer $SPICY_API_KEY" \
-H "Content-Type: application/json" \
-H "Idempotency-Key: $SPICY_IDEMPOTENCY_KEY" \
-d '{
"model": "alibaba/wan-2.7/text-to-image",
"input": {
"prompt": "A ceramic teapot on a linen cloth, soft window light, muted palette",
"resolution": "1k",
"num_outputs": 1,
"aspect_ratio": "1:1"
}
}'
# Copy data.taskId from the response into TASK_ID, then poll the status.
TASK_ID="TASK_ID_FROM_RESPONSE"
curl "https://api.spicyapi.ai/api/v1/jobs/recordInfo?taskId=$TASK_ID" \
-H "Authorization: Bearer $SPICY_API_KEY"
See the API reference for authentication, the complete schema, polling, and error handling.
