Wan 2.7

Video GenerationReference GuidedVideo Editing

Wan 2.7 Video on APIXO brings four Alibaba video workflows into one operational route: prompt-led creation, frame-guided animation, multimodal reference generation, and instruction-based video editing. Produce 720p or 1080p clips with mode-specific visual, audio, duration, and framing controls for advertising, narrative development, localization, and post-production pipelines.

Modes

Text · Image · Reference · Edit

APIXO Price

From $0.10 / SecBackend-priced per-second video route

Resolution

720p · 1080p

Output Duration

Up to 15 Seconds

Delivery

MP4 URL · Async

Loading workspace...

Capabilities

Shape motion, subjects, sound, and existing footage

Prompt-Led Scenes

Text-to-video turns a Chinese or English description into a 2–15-second clip. Define subjects, action, camera behavior, scene progression, and sound direction in a prompt of up to 5,000 characters.

Frame-Guided Motion

Image-to-video can animate one first frame, connect first and last frames, or continue a source video with an optional final-frame target. A prompt is optional when the supplied visual already establishes the scene.

Multimodal References

Reference-to-video combines as many as five ordered image and video references. Use them to establish subjects, appearance, motion, or scene context without treating every reference as a fixed output frame.

Slot-Aligned Voices

Reference requests can associate public audio files with individual reference slots. APIXO orders image slots before video slots and accepts empty placeholders when a later reference needs audio but an earlier one does not.

Instruction-Based Editing

Video-edit applies written transformation instructions to one existing 2–10-second clip and can incorporate up to four supporting images. It is suited to style changes and guided visual revisions rather than timeline-based manual editing.

Audiovisual Generation

Wan 2.7 supports synchronized audiovisual output and optional custom audio in compatible generation workflows. Multi-shot structure can be described through prompt wording, but APIXO does not expose a dedicated `shot_type` parameter on this route.

Route Details

Wan 2.7 Video specifications on APIXO

Mode-specific inputs, duration limits, and delivery rules for the operational APIXO route.

720p / 1080p

Resolution

16:9 / 9:16 / 1:1 / 4:3 / 3:4

Aspect Ratios

2–15 Seconds

Text / Image

2–15 Seconds · 10s with Video

Reference

0 or 2–10 Seconds

Video Edit

Polling or Callback

Delivery

Production Uses

Build distinct video workflows around the right source material

Creative Development

Narrative Concept Clips

Translate treatments and shot descriptions into short audiovisual scenes for pitch development, story exploration, and previsualization. Multi-shot prompting helps communicate a progression of actions without requiring source footage at the concept stage.

Controlled Animation

Product and Artwork Motion

Animate a supplied product composition, key visual, illustration, or opening frame. Add an optional last frame when the ending composition matters, while reviewing fine geometry, embedded text, and brand details before publication.

Subject Direction

Reference-Led Campaign Scenes

Combine subject images, performance footage, and slot-aligned voice references to develop campaign scenes around recurring visual elements. Complex identities, interactions, speech, and rapid movement may still require several generations and human review.

Post-Production

Guided Footage Restyling

Submit a short source clip with a transformation brief and optional visual references to explore new treatments, environments, or art directions. Choose whether Wan decides how audio should be handled or preserves the original soundtrack.

Notes & FAQ

Mode-specific requirements and billing

Integration Notes

Text-to-video and reference-to-video require a prompt; APIXO’s current detailed route documentation also requires a non-empty prompt for video-edit.

Image-to-video accepts one or two frame images, or one 2–10-second continuation video with no more than one optional last-frame image.

Reference-to-video requires 1–5 combined image and video references; adding any video reference reduces maximum output duration to 10 seconds.

Video-edit requires exactly one public MP4 or MOV source clip between 2 and 10 seconds and supports up to four optional reference images.

Aspect ratio is unavailable for image-to-video because its framing follows the supplied source; other modes apply their documented ratio rules.

A fixed seed can improve reproducibility but does not guarantee identical output, and enabling watermark adds the fixed bottom-right text “AI 生成.”

Frequently Asked Questions

The route exposes text-to-video, image-to-video, reference-to-video, and video-edit. These are separate workflows with different prompt, media, ratio, duration, audio, and billing rules; fields valid in one mode should not be assumed valid in another.

Text-to-video, image-to-video, and video-edit cost $0.10 per billable second at 720p or $0.15 at 1080p. Reference-to-video costs $0.15 per second at 720p or $0.24 at 1080p.

APIXO charges video-edit for the capped input duration plus the requested output duration. With duration=0, output follows the source duration and billable seconds equal twice the capped input duration. With a 2–10-second output request, APIXO adds that duration to the capped input duration.

audio_setting belongs only to video-edit. auto lets the model decide whether prompt sound instructions call for regenerated audio and may otherwise retain the source audio. origin forces preservation of the input audio and skips regeneration.

APIXO maps audio_urls against reference slots in order, with all image references first and video references second. Use an empty string to skip one slot without shifting later audio. Each non-empty entry acts as voice guidance for its associated reference.