Video to Video

• Duration: 5, 10, 15, 20, 25, or 30s

• Resolution: 480p / 720p / 1080p

• Best for: prompt-driven video-to-video generation with reference videos (up to 3), reference images (up to 9), and reference audio - building long, multi-shot edits or transformations up to 30 seconds, with optional native audio written by the model. Trained on a massive labeled dataset of NSFW content.

• Strengths: can write its own soundtrack via the Generate Audio toggle, follows long screenplay-style prompts broken into timestamped shots, reference tokens ([img1], [vid1], [audio1]) let you address exactly which upload to use, stronger multi-subject consistency and camera control than earlier Wan models.

• Trade-offs: best results need a prompt matched to the duration - short prompts on long durations hand plot and shot choices to the model. No negative prompt; unwanted elements must be described explicitly.

• Content notes: Filters: Minimal

• Duration: 5, 10, 15, 20, 25, or 30s

• Resolution: 480p / 720p / 1080p

• Best for: the same prompt-driven video-to-video generation as Wan 3.0 Spicy — reference videos (up to 3), reference images (up to 9), and reference audio — but returned about 40% faster.

• Strengths: identical output quality and feature set to Wan 3.0 Spicy (Generate Audio toggle, screenplay-style prompts, reference tokens, multi-subject consistency); the difference is turnaround time, which matters when testing ideas or iterating on a shot.

• Trade-offs: costs more per second than Wan 3.0 Spicy for the same result; same prompt-length sensitivity and lack of negative prompt.

• Content notes: Filters: Minimal

• Duration: 5–15s

• Resolution: 720p / 480p

• Best for: prompt-driven video generation with first/last frame and multimodal reference (images, videos, audio) — editing a finished video by replacing a costume, background, lighting, or setting while preserving the subject's movement, timing, and identity. Trained on a massive, annotated dataset of NSFW content.

• Strengths: a clear distinction between "what to keep" and "what to change" in the prompt yields predictable results without "rebuilding" the scene. Preserves the original animation and body proportions well when the prompt is precisely worded. Handles references [video 1]/[audio 1]/[image 1] accurately and can generate audio when no reference is provided.

• Trade-offs: the order in which elements are described is critical — if you start with the changes rather than what to keep, the model may lose the original movement and timing. The original video reference [Video 1] must be clear for reliable results. Max length of the 3 video references is 15 seconds. Slower than Seedance 2.0 Fast Spicy.

• Content notes: Filters: Minimal

• Duration: 5–15s

• Resolution: 720p / 480p

• Best for: faster, cheaper video generation with the same first/last frame and multimodal reference approach as Seedance 2.0 Uncensored — good quality, editing pre-recorded videos (costumes, backgrounds, lighting) while preserving movement, timing, and facial identity. Plus additional training on a massive labeled dataset of NSFW content.

• Strengths: fast processing time; clearly distinguishing between "what to keep" and "what to change" in the prompt yields predictable results without having to "rebuild" the scene. Faster and cheaper than Seedance 2.0 Uncensored.

• Trade-offs: the order in which changes are described is critical — if you start with the changes instead of what to keep, the model may lose the original motion and timing. The original video reference [Video 1] must be clear. Max length of the 3 video references is 15 seconds. Quality is lower than Seedance 2.0 Uncensored.

• Content notes: Filters: Minimal

• Duration: 5, 10, 15, 20, or 30s

• Resolution: 480p / 720p / 1080p

• Best for: prompt-driven video-to-video generation from up to 10 reference videos, 30 reference images, and 10 reference audio clips — addressed in the prompt as [vid1], [img1], [audio1] — into long, multi-shot results up to 30 seconds. Trained on a massive labeled dataset of NSFW content.

• Strengths: more natural, less "AI-looking" footage — improved camera transitions, subject stability, lighting, skin, and eye detail; timestamped prompts ("0–5s: …, 6–10s: …") control what happens and when; optional native audio via the Generate Audio toggle.

• Trade-offs: the priciest Seedance model — 1080p costs about 3.4× the 720p rate, and the seconds of your reference videos are billed along with the result, so test at 480p/720p first. No negative prompt. The video may come without audio even with Generate Audio on.

• Content notes: Filters: Minimal

• Duration: up to 30s

• Resolution: 720p / 480p

• Best for: bringing a still character image to life by applying the exact body movements, gestures, facial expressions, and lip movements from a reference video. Ideal for character performances, dance sequences, expressive animations, and multi-character scenes, plus additional training on a massive labeled dataset of NSFW content.

• Strengths: excellent motion fidelity and semantic understanding of movements, high identity and background consistency, strong performance with non-human and stylized characters, natural expressions, stable output across complex actions. The most stable option for bringing a photo to life.

• Trade-offs: best results with well-matched composition, clear reference video (good lighting, minimal occlusion), and compatible character image.

• Content notes: Filters: Minimal

• Duration: up to 30s

• Resolution: 720p / 480p

• Best for: bringing a still character image (portrait, illustration, or design) to life by applying the exact body movements, gestures, and facial expressions from a reference video. Ideal for character performances, dance sequences, and expressive animations, plus additional training on a massive labeled dataset of NSFW content.

• Strengths: strong motion fidelity with natural expressions, high identity consistency, seamless transfer of complex actions.

• Trade-offs: best results with well-matched composition, clear reference video (good lighting, minimal occlusion), and compatible character image. May be unstable — see the note below.

• Content notes: Filters: Minimal

• Duration: up to 30s

• Resolution: 720p / 480p

• Best for: replacing the main character in an existing video with a subject from your image while preserving the original motion, scene, lighting, and camera work. Perfect for character swaps in footage, presenter replacements, and seamless integration, plus additional training on a massive labeled dataset of NSFW content.

• Strengths: maintains original scene elements and motion, strong identity preservation, realistic integration with environment and lighting.

• Trade-offs: best results with good pose/composition matching between image and video, clean reference footage. May be unstable — see the note below.

• Content notes: Filters: Minimal

• Duration: up to 30s — longer videos are automatically clipped

• Resolution: 1080p only

• Best for: applying choreography and movements from a reference video to a subject in a character image. Ideal for complex dance moves, gestures, and full-body performances. Great for dance sequences, motion transfer, and creating consistent character animations.

• Strengths: extracts and transfers full motion/choreography (including complex actions, hands, facial expressions) from a reference video. Keep-original-sound option. Strong motion fidelity and character consistency. Highest resolution of any model here.

• Trade-offs: requires good reference video quality (clear, unobstructed movements) and a compatible character image. Stricter safety filters than the other models.

• Content notes: Filters: On

• Duration: 5, 6, 8, 10, 12, or 15s

• Resolution: 480p / 720p / 1080p

• Best for: reworking an existing video with a text prompt, optionally combining it with reference images, extra videos, and audio (up to 9 images, 3 videos, 3 audio clips, 12 files total) to steer the result.

• Strengths: reference tokens (Image 1, Video 1, Audio 1) let you point to exactly which upload drives which part of the edit; native stereo audio on every result, no toggle needed; flexible aspect ratios (16:9, 9:16, 4:3, 3:4, 1:1, 21:9, or Adaptive).

• Trade-offs: billed for both the source video's seconds and the output's seconds. No negative prompt; describe exclusions at the end of the main prompt. Audio can't be disabled. Long prompts are compressed internally, so shorter, precise wording tends to work better.

• Content notes: Filters: On — SFW only, does not support NSFW content.

You'll receive an error — but your credits will be automatically refunded.

Yes, up to 30 seconds.

This can happen when the face isn't clearly visible in all frames, the video contains fast motion or cuts, or lighting changes significantly. For best results, use stable videos with consistent lighting and a visible face.

Your task may fail due to unsupported or corrupted files, very low-quality inputs, content restrictions depending on the selected model, or unstable motion in the source video. Try simplifying your inputs and using a shorter or clearer video.

Animate follows the motion of the video and is faster but less stable. Kling 2.6 Motion provides higher quality and consistency, but uses stricter safety filters.

Kling 2.6 Motion supports a maximum duration of 30 seconds. If your video is longer, it will be automatically clipped.

Yes, you can paste links directly into the tool. Links work only with some models, and Seedance, Wan 3.0, and MiniMax H3 require an uploaded video file. Make sure the video is public and accessible.

No — credits for failed tasks are automatically refunded to your balance.