Wan 2.2 Video Generator｜Photo to Clip
Upload one photo or type one sentence and get a 5–15s cinematic clip. No installs, phone-friendly, newer Wan models inside.
Wan 2.2-style clips free online: one photo or sentence becomes a 5–15s clip. No downloads, phone-ready, faces locked. Try the Wan 2.2 video generator free.
Features
- One photo is enough——Start frame is optional. Upload one photo for image-to-video, or type one sentence for text-to-video — both land the same cinematic Wan look.
- Identity stays locked——Your upload is locked as the first frame and the face is never re-described. Side angles plus small motions survive best.
- 5–15s, ready to post——5s, 10s or 15s; landscape, vertical or square. Honest note: the standard workflow is silent — no native audio.
- No installs, phone-ready——No 4090, no ComfyUI, no downloads. Runs in the browser, phone-friendly, driven by newer Wan-series models.
How it works
- Upload a start frame Selfie, pet, product or landscape — sharp and unobstructed works best. Skip it to go pure text-to-video.
- Describe it, pick specs Write subject + action + camera + light, then pick 5/10/15s, aspect and style. Presets below fill it for you.
- Take the clip, post it Get a 5–15s silent clip and post it straight to TikTok, Reels or the group chat.
FAQ
- Is this the original Wan 2.2 model?
- Honestly: no. Loomi does not host the original Wan 2.2 weights, so this page is powered by newer models of the same Wan series (wan-3.0 primary, wan-2.7 fallback) covering all core 2.2 abilities: text/image-to-video, 720p, 5–15s. If you need the exact 2.2 weights for local deployment, get them from the official GitHub repo Wan-Video/Wan2.2.
- Can one photo really become a video?
- Yes. That is image-to-video: your photo becomes the first frame and the model animates it. It does not have to be front-facing — side profiles, full body, pets, products and landscapes all work. Use a sharp, unobstructed photo; for turns and walks a side-angle photo actually glitches less.
- How do I keep it realistic, no glitches?
- Three rules: a good first frame, restrained motion, one camera move. Describe only action + camera + light (e.g. slow push-in, natural blink, golden hour), never re-describe the face. Keep hands away from close-ups, or re-roll the shot separately.
- Do the videos come with sound?
- No. The standard Wan 2.2 workflow has no native audio — clips come out silent, and we will not promise voiceovers, dialogue or lip-sync (that is the separate S2V capability). Add music or narration afterwards in any editor.
- Can I play on my phone? Any download?
- Yes. Everything runs in the browser on Loomi — no downloads, no 4090, no ComfyUI workflow. Upload from your phone and take the 5–15s clip straight to TikTok or the group chat.
- Is it free? Do I have to wait in line?
- Log in to Loomi and generate directly — usage costs credits per clip, no hardware to buy and no local queue. ~5s at 720p is the sweet spot for a first try.
- Can I use the clips commercially?
- The Wan 2.2 weights are open source under Apache 2.0. For clips you make on Loomi, commercial use follows Loomi's platform terms — and if a real person's face is in the frame, get their permission first.
Try it free