alibaba/wan-3-video
Wan 3.0 video — one model covering text-to-video, first-frame, first+last-frame and reference-based generation; the mode follows the media you send. prompt (required unless media is given); first_frame_url; last_frame_url (use both to render between two stills); reference_image_urls (up to 10); reference_video_urls (up to 5 clips, 15s total); reference_audio_urls (up to 5); resolution 480p|720p|1080p (def 1080p); aspect_ratio adaptive|16:9|4:3|1:1|3:4|9:16; duration whole seconds 2-30 (def 5) — longer than any other model here; audio (def true); negative_prompt; seed; watermark.
Served via the relay: POST /v1/alibaba/generations, poll GET /v1/alibaba/generations/{task_id}. Per-second pricing by resolution, billed by the relay from the vendor's reported usage.
| Slug | alibaba/wan-3-video |
| Kind | video |
| Vendor | alibaba |
| Endpoint | POST /v1/alibaba/generations |
Charged per job from what the generator reports it rendered — your Logs show the exact figure for each one; see balance and billing.
Calling it
curl https://api.sociaro.com/v1/alibaba/generations \
-H "Authorization: Bearer $SOCIARO_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "alibaba/wan-3-video",
"prompt": "a paper boat on a rain-soaked street",
"resolution": "720p",
"duration": 5
}'
# then poll the job id it returns:
curl https://api.sociaro.com/v1/alibaba/generations/JOB_ID \
-H "Authorization: Bearer $SOCIARO_API_KEY"Parameters
Types, ranges, defaults and what each field does — the list below is the names only.
promptfirst_frame_urllast_frame_urlreference_image_urlsreference_video_urlsreference_audio_urlsresolutionaspect_ratiodurationnegative_promptseedprompt_extendwatermarkaudio
Anything not listed here is refused rather than ignored, so a typo fails loudly instead of quietly producing something else.