wan 3.0 ai video generator

wan-3.run is an independent, third-party interface for Wan 3.0 - the video generation model Alibaba's Tongyi Lab released in August 2026. It runs in a browser, with no Alibaba Cloud account, no region and key setup, and no local GPU. Four inputs reach the same model. A text prompt of up to 20,000 characters. A photo used as the first frame, the last frame, or both. A reference set of up to 10 images, 5 video clips and 5 audio tracks, addressed by number inside the prompt so one face, one product or one voice holds across a whole take. Or a document - a PPT, PDF, spreadsheet or public web page up to 50 pages - which Wan 3.0 reads and turns into new footage rather than narrating your slides. You can also rewrite footage you already have without regenerating it. Output is 2 to 30 whole seconds in a single pass at 30 fps, at 480P, 720P or 1080P, with speech, effects and music generated alongside the picture. Turning sound off does not make it cheaper, because it is one pass. The free clip runs wan3.0-video - the same model ID every paid plan runs, not a cheaper substitute engine - and every result prints the model, the resolution, the length and the upstream task ID. Failed and expired jobs are refunded in credits automatically, with no ticket. Plans are $9.90, $29.90 and $79.90 a month billed yearly, and credits burn at the same per-second rate on every plan.