Alibaba’s Wan team is hosting its open-source AI video tool online, offering easy access to people who can't or don't want to host the free model locally. Wan is a powerful model often leading the VBench leaderboard, outperforming both open-source and commercial competitors.
We are listing this platform separately for clarity purposes, if you're looking for a free PC-based Wan, please check out this entry. Here we're highlishting what the online platform offers.
In August 2025 Wan adds their avatar lip sync tool. Wan2.2-S2V is a 14B model made for cinematic, audio-led human animation. It goes past simple talking heads and aims for pro quality used in movies, shows and web stuff. (It’s open-source too.) It keeps movement consistent across long videos. You can guide motion and background with clear instructions.
Educators and TrainersCreative ProfessionalsContent CreatorsMedia and Film MakersMarketing and Branding SpecialistsDevelopers and Tech CreatorsNonprofit and Advocacy CreatorsSmall Business OwnersEntertainment and Performance ArtistsProfessional Content Creators
A wide shot opens on a soaked, winding asphalt road glistening under a low mist. A lone adult sheep stands in the center of the frame, its damp wool clinging to its body as the sky hangs heavy with haze. The camera performs a double dolly zoom, pulling backward on the dolly as the lens tightens its focal length, causing the road and tree line behind the sheep to stretch and compress while the sheep remains visually stable in the center of the frame. The sheep lifts its head toward the camera with a slow, curious gaze.
A lightning strike erupts behind the sheep, illuminating drifting sheets of fog. A swirling burst of dust rises around its legs, wrapping its body in a violent coil of wind and fog. The sheep’s form enlarges as coarse grey fur forces through its skin, limbs extending, its jaw reshaping into a snout filled with sharp teeth as it fully transforms into a roaring grey werewolf.
The camera shifts into a handheld tracking shot, following the werewolf from the front as it charges forward on the wet road. Its red eyes lock directly into the lens while its mouth opens wide, revealing rows of wet, jagged teeth, illuminated intermittently by flashes of lightning.
“Eat Something Else” and "Bad marketing by www.aicreators.tools". The text remains on the package throughout the scene.
Beams of light move dramatically over the product.
Suddenly, the bag wobbles, then bursts open near the top and right corner, sending misshapen rotten with spots of mould cookies flying out of the opening through the air. Each cookie has a melted chocolate face — frowning, screaming, or looking dizzy and confused as they spin and tumble in exaggerated animation around and above the blurred-out ripped-open package. Their faces are animated.
The camera captures the flying cookies in crisp detail as one spins close to the lens, its face frozen mid-scream, while another’s expression subtly melts and slides down its surface. The tone is darkly funny — a parody of an overproduced snack commercial gone horribly, cinematically wrong. Backdrop features a sleek studio set with soft gray tones and a glossy floor reflecting bright overhead spotlights, wall having blurred text "www.aicreators.tools". More closed blurred cookie packages line the background. Cinematic lighting, glossy reflections, absurd humor, parody-commercial aesthetic.
This was fun. But had to tweak the prompt a bit to prevent text to detach from packaging. Smaller font still does for a short while. This is sound-driven gen, speech by ElevenLabs.
Fat man skates with wild enthusiasm. He twirls, leaps, and slides across the frozen lake in chaotic elegance — the ice groaning dramatically beneath him as shards spray with every spin. His face is locked in fierce concentration, veins bulging, as he attempts a flurry of ballerina pirouettes that verge on both impressive and catastrophic. The camera whips between angles — darting from low gliding shots near his skates to fast, handheld-style closeups of his flushed expression, then swinging wide to capture his flailing, magnificent form against the sweeping winter landscape. The rhythm of the motion feels erratic, almost like a fever dream, as snow whirls violently around him.
High-quality, cinematic footage. Two dogs (a Miniature Schnauzer on the left and a chocolate-brown Labrador on the right) sit in a professional podcast studio'
On the walls, there are framed pictures of various dogs with vane and noble expressions and a large certificate-award reading "Certified Good Boy*" in bold and below, smaller on right "*Self-Certified"
Dim studio lighting from multiple spotlights.
The Miniature Schnauzer starts by saying enthusiastically: “Mine talks to the TV like it can hear him.”
The Labrador replies: “Mine argues with it. ARGUES. With a rectangle.” In the end both dogs laugh together, creating a humorous and witty atmosphere.
A man runs towards the camera in a park. As the character runs past, the camera performs a fast whip pan to follow, blurring the entire scene into horizontal streaks. The blur acts as a transition: when the whip pan stops, the scene is now a city and the man is running away from the camera on a busy sidewalk.
A confident 25-year-old woman stands before a vibrant yellow wall painted with the phrase “AI creators tools.” She wears a chic purple jacket and embellished jeans, a black backpack in her hand. In a calm, rhythmic motion, she begins modeling her outfit with poise — stepping closer in fashion-forward pose while looking confidently at the camera, projecting strength, then reatreating and adjusting her jacket, turning around herself. The camera remains almost still, holding focus on her, while a gentle parallax effect and soft depth-of-field blur shift subtly as she moves. The diffused lighting glows warmly, reflecting delicate highlights off her clothes, giving the entire frame a rich, fashion-editorial quality in crisp 4K.
Close-up of a purple-yellow slide falling onto a semi-reflective floor, followed by another slide creating a pair. The camera tracks closely as bare feminine feet enter the frame, putting slides on one by one. It slowly zooms out, capturing the woman now wearing the slides walking from left to right, taking three steps before stopping and turning towards the camera. We see her legs and waist in smart loose red cotton shorts. Camera then quickly pulls back to reveal the whole woman who is wearing a grey t-shirt with dark-grey shimmering collar and sleeve details tucked into the red shorts. She is a brunette with shoulder-length hair, smiling
Wan is struggling with this one. Even when it does the beginning right, there are 2 extra slides. Tried 'putting/placing' slides on variation, tried adding 'Final shot is the woman wearing the slides smiling, the floor around her is clean having no other objects' - simply makes extra footware disappear in the end but doesnt prevent it appearing.
Starts with a close-up of a single purple-yellow slide resting upright on a semi-reflective floor, identical to the one shown. Another matching slide drops gently beside it, completing the pair. After a brief pause, two bare feminine feet enter the frame from above and step into the slides one by one, clearly showing the motion of putting them on. The camera follows this action smoothly, then slowly zooms out as the person walks from left to right, taking three relaxed steps before stopping and turning toward the camera. She’s wearing smart loose red cotton shorts and a grey t-shirt with dark-grey shimmering collar and sleeve details tucked in. Shoulder-length brunette hair, smiling cutely. Artistic brushstrokes appear behind her, text spelling "AI CREATORS TOOLS" Cinematic 4K detail.
Hip-hop dance, two cute anthropomorphic cats in the frame are dancing Hip-hop in sync staying close side-by-side. Lights animated, subtle fog behind, groovy hip-hop electronic beat playing. mood frolic and energetic
[00:00-02.00] cats are dancing while facing the viewer
[02:01-03.00] cats swirl around once
[03:01-05.00] cats are dancing turned to the side but looking at the viewer with their heads turned towards the camera
Wan ignores timestamps. Does all the moves requested in prompt, except cats swirls quickly more than once. And video ends with them turning their backs to the camera. Dropped the camera motion for this prompt. With it, the dancers get out of sync with each other - only happens when prompts call for dancers turning around in sync, no issue if they keep facing same direction while dancing.
Like some other models, Wan 2.5 loses character reference when going form painting to realistic, even after tweaking the prompt to say more clearly 'Reimagine the subject of THIS painting'.
Wan2.5 doesn't really do jump cuts? Judging form 3 different prompts and ~7 generations. The character will turn around and run a different direction if a prompt calls to see them from another angle.
Wan2.5-Preview is now out.
It runs on a native multimodal design that works across text, images, video and audio. It can generate videos with synced audio covering vocals, sound effects and background music. It can follow directions more clearly to produce photorealistic results, varied art styles, imaginative text effects and pro-level charts.
model
August 27, 2025
Avatar lip sync model Wan2.2-S2V-14B is out and available for use on the platform.
model
May 17, 2025
Five days left to take advantage of Wan's online platform membership sake (till May 23). Get 50% off plus 1 month free if you sign up for yearly plan, which starts as low as $60.
promo
Useful Links
No additional links available for this tool.
This page was last updated on October 7, 2025 at 3:42 AM