Wan 2.7 Image is Alibaba's image model for generation and editing in one endpoint. Write a prompt for text-to-image, pass reference images to edit or fuse, box a region for precise interactive edits, or batch a consistent series — up to 4K on the Pro tier, all through reApi's unified API.
Up to 5000 characters · required for text-to-image · optional (but recommended) when editing
Public HTTPS URLs only (no base64) · up to 9 · output ratio follows the last image
1–4 normally · up to 12 in group mode · billed per generated image
1:1, 16:9, 9:16, 4:3, 3:4, 3:2, 2:3 — pair with the resolution tier
1K / 2K everywhere · 4K is Pro text-to-image only
e.g. blurry, distorted, low quality
Off by default
0–2147483647 · same seed + params gives a similar look
On by default · only active for text-to-image with no reference and no group mode
Raises the count cap to 12 · best for storyboards and comics
JSON: [[[x1,y1,x2,y2], …≤2], …] — one entry per reference image, [] to skip
JSON: 3–10 hex/ratio items summing to 100% · non-group mode only
Uncensored generation is available via the API. Use your API key to generate; browser generation is disabled.
Try one of these prompts
Credit-based — 1 credit = $0.001 USD. Pay only for completed generations.
One sample at the cheapest tier (1 image).
Wan 2.7 Image is Alibaba's image generation and editing model, exposed on reAPI as one async endpoint in two tiers — Standard at 1K/2K and Pro, which adds 4K for text-to-image. The same request handles text-to-image, editing and fusion from up to 9 reference images, box-targeted local edits, a custom colour palette, and sequential batches of up to 12 images across 7 aspect ratios, billed at a flat rate per image regardless of resolution.
Real-world workflows and production use cases you can build and ship with this model.
Wan 2.7 Image is a single endpoint for two jobs. Send a prompt and it renders text-to-image; add reference images and the same call switches to editing — restyle a product shot, swap a background, fuse elements from several photos into one scene. There is no separate edit model to learn: the request shape is identical, the mode is implicit in whether you attach images, and the output ratio follows your last reference.
Open the playground
Pass bbox_list and Wan 2.7 Image edits exactly where you point. Box the clock on the desk, the logo on the car, the empty wall — coordinates are absolute pixels on the source image, up to two regions each — and the model composites the change with matched lighting and perspective. It is interactive inpainting through the API, no mask painting required.
Turn on group mode and Wan 2.7 Image returns a connected set — same character, same style — across up to twelve frames in one call. It is built for storyboards, comic panels, and seasonal campaign sets where continuity matters. One request, one task ID, a coherent series back, with no outer loop to script or stitch.
Text-to-image, image editing, multi-image reference, interactive region edits, and group series all run through the same request to /api/v1/images/generations. Pick the tier with the model id — wan2.7-image or wan2.7-image-pro — and let the inputs decide the mode. No branching client code, no second model to integrate.
1 credit equals $0.001 USD. Wan 2.7 Image bills a flat rate per successfully generated image — independent of resolution and aspect ratio — so cost is a single multiply on your count. The Pro tier renders text-to-image up to 4K when you need print- and retina-grade detail; failed and rejected requests are never charged.
Submit a Wan 2.7 Image task, get a task_id back, poll until it completes. It drops into the same task plumbing as every other reApi model, and the JSON matches the OpenAI image-generations contract — so adding Wan 2.7 Image is a model-id change, not a new integration.
Both generate and edit images from one API. Wan 2.7 Image leans into editing control — interactive region edits, multi-image fusion, consistent group series — with up to 4K and flat per-image pricing. Here is how the two compare on publicly documented behavior.
Comparison reflects publicly documented behavior at the time of writing. Model behavior and pricing can change; check the pricing card above and the API docs for current values.
Sign up at reApi and grab a key. Free signup credits cover dozens of Wan 2.7 Image generations — enough to ship a first hero shot or edit before topping up.
OpenPOST to /api/v1/images/generations with model: wan2.7-image-pro (or wan2.7-image), a prompt, and optional size / resolution / image_urls. Wan 2.7 Image returns a task_id immediately.
OpenGET /api/v1/tasks/:id until status is completed. The Wan 2.7 Image response carries image URLs; mirror them to your own storage if you need them long term.
Opencurl -X POST https://reapi.ai/api/v1/images/generations \
-H "Authorization: Bearer YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "wan2.7-image-pro",
"prompt": "a flower shop with delicate windows and a wooden door",
"size": "16:9",
"resolution": "2K"
}'Rejected synchronously with the failing field named. Check enum values, ranges, and URL fields — nothing is charged.
401 means a missing or invalid Bearer key; 402 means the reserve exceeds your balance. Manage keys and credits in the console.
A provider may reject unsupported or invalid inputs. Check the task error and media requirements before retrying. Failed tasks are refunded.
A task that reaches a failed state is never charged — the reserve refunds automatically. Keep the task id and retry when ready.
Common questions about this model.
Configure parameters in the Playground, then use your API key to generate.
