CREATION MODES

MODELS

Luma Midjourney OmniHuman InfiniteTalk Topaz Upscale Lip Sync

Guest trial

Attempts left 3 Available

My Library

· 0 videos

Your next video starts here

Finished and in-progress generations will stay organized in this Account Library.

Write a prompt →

AI video model workspace

Wan AI video generator for text, speech, references, and edits

Use Wan on ReelPush for text, image, speech-driven animation, reference-to-video, or source-video editing through separate variants.

No credit card Account-owned Library

17available models

4creation modes

1Account Library

An AI video maker for teams that publish often

DTC brands Creative agencies In-house teams Independent creators

See what reelpush.ai creates

A precision product reveal on a reflective studio surface

reelpush.ai

A vertical skincare routine with clean morning light

reelpush.ai

Aerial movement over a quiet coastline at sunrise

reelpush.ai

Macro coffee pour with warm, tactile detail

reelpush.ai

A bold sneaker launch built for a 9:16 feed

reelpush.ai

A cinematic night market with a slow camera push

reelpush.ai

Soft fabric movement for a seasonal apparel drop

reelpush.ai

A compact unboxing sequence with crisp transitions

reelpush.ai

Wan AI video generator

Wan on ReelPush

Wan is Alibaba's broad AI video model family. Its ReelPush catalog spans generation, animation, speech, reference, and editing tasks rather than one interchangeable prompt box.

Alibaba

Wan capabilities available now

15 available variants, text, image, and video input, up to 580p, 1 to 15 seconds, 9:16, audio-capable variants, source image, source video, and up to 5 reference images, 5 reference videos, and 1 reference audio

Route speech to a speech model

Wan speech variants pair a source image with audio reference instead of treating dialogue as a plain text-generation request.

Edit the clip you already shot

Wan Edit Video variants require source video and can transform footage without rebuilding the scene from zero.

Use references across media

Wan 2.7 R2V can accept image, video, and audio references for a brief that needs several kinds of evidence.

ReelPush workflow

How to use Wan on ReelPush

Inventory the source material

Identify whether the brief starts with text, an image, speech audio, references, or an existing video.

Open the matching Wan variant

Choose the catalog entry whose required inputs match that source instead of forcing media into a general generator.

Verify the transformed asset

Review identity, timing, inherited details, audio, and unwanted changes before exporting the saved result.

Short-form video uses

Three jobs suited to Wan

Speech-driven character insert

Pair one subject image with prepared audio for a short dialogue asset.

Source-footage style revision

Send an existing product clip through a Wan edit variant with a focused transformation prompt.

Mixed-reference campaign scene

Use a reference-capable variant to coordinate a product, motion cue, and sound source.

Choose by task

When to compare Wan with another model

Wan or Kling

Both have broad catalogs. Wan includes explicit speech and source-video editing paths; Kling includes motion-control and avatar families.

Kling Video Generator

Wan or HappyHorse

Compare reference requirements and editing controls. HappyHorse has a focused character-reference route plus a three-second video edit variant.

HappyHorse Video Generator

Evidence

Product facts and independent material

Capability facts come from ReelPush's live model catalog. External links provide workflow context and are not endorsements or customer testimonials.

ReelPush Wan capability catalog Independent Wan workflow video

Product boundaries

Generate a video asset, then complete the publishing work

ReelPush generates and stores video assets from the selected model's supported inputs. It does not automatically add final captions, verify claims, clear likeness or music rights, schedule posts, or publish to a channel. Review every output for identity, objects, movement, audio, brand rules, consent, and platform requirements before use.

Model questions

Wan questions, answered

Which Wan workflow should I start with?

Start with the source you already trust. Use text for an open brief, image for a controlled starting frame, speech for portrait plus audio, animation for movement evidence, reference-to-video for mixed guidance, or Edit Video when an existing clip must remain the foundation.

Can Wan edit an uploaded video?

Yes. Enabled Wan Edit Video variants require a source video and expose their own resolution, duration, aspect-ratio, and audio behavior. Upload a short supported clip, describe one deliberate change, and inspect which original details were retained or altered in the result.

Does Wan accept reference audio?

Selected Wan variants accept audio references. The speech workflow uses audio to drive a source image, while a mixed-reference variant can accept audio alongside visual references. The editor labels the active contract, limit, and required media so these uses remain distinct.

What Wan resolutions are available?

Current enabled entries range across 480p, 580p, 720p, and 1080p depending on the variant and task. ReelPush only offers combinations declared by that catalog entry. Select the workflow first, then compare its resolution and duration controls with the final delivery need.

Unlock this generation

My Library

Your next video starts here

Generating video