Increase video dimensions by 2x with a single input clip. Process footage from 0.5 to 600 seconds for larger previews, presentation assets and video delivery workflows.
Models
All Models
Capture a person's video performance as 3D skeletal animation. Export GLB or FBX with rig options, frame-rate control and optional smoothing for animation workflows.
Animate a character image with motion from a driving video. Keep the source background, choose a solid color, or export an alpha mask for flexible video compositing.
Create 3D skeletal animation from a motion description. Export GLB or FBX with Mixamo or MetaHuman rigs, adjustable duration, prompt guidance and optional joint smoothing.
Create images from text with up to 8 optional visual references. Explore 7 aspect ratios, 1K or 2K output, and seeded variations for designs, characters and campaign art.
Turn a full-body character image into a rigged .vsplat asset for Viggle Splat Engine in Unity. Choose joint sets, pruning options and an optional PNG thumbnail preview.
Replace people in a driving video with 1 to 5 character images. Preserve the performance and timing while exploring new cast designs for short clips and character tests.
Animate an image with adjustable motion speed from 0.025x to 1x. Generate up to 121 frames at 24 fps for controlled action studies, product shots and expressive motion.
Restore damaged archival footage in natural color. Use a reference clip 2.04 to 4.33 seconds long, a scene-specific prompt and a seed to explore controlled video transformations.
Refine fine texture and detail in an existing video. Use a reference clip up to 10.33 seconds long, a scene-specific prompt and a seed to explore controlled video transformations.
Turn a layout or playblast into a finished-looking shot. Transform short reference clips with scene-specific prompts and a seed for reproducible tests and production previews.
Animate characters, props and settings from a reference sheet. Transform reference clips with scene-specific prompts and a seed for reproducible tests and production previews.
Add moving water to a dry reference scene. Use a reference clip 4.04 to 8 seconds long, a scene-specific prompt and a seed to explore controlled video transformations.
Upscale a video to twice its input dimensions. Use a reference clip up to 10.33 seconds long, a scene-specific prompt and a seed to explore controlled video transformations.
Ask NVIDIA Nemotron 3.5 Super about a video: summaries, scene descriptions, action and event analysis, with an optional reasoning mode.
Turn a still image into a cinemagraph with selective motion and a fixed camera. Generate up to 121 frames at 24 fps for product loops, ambient visuals and social posts.
Ask NVIDIA Nemotron 3.5 Super about an image: descriptions, text reading, visual question answering, with an optional reasoning mode.
Add moving water to a dry reference scene. Use a reference clip 4.04 to 8 seconds long, a scene-specific prompt and a seed to explore controlled video transformations.
Upscale a video to twice its input dimensions. Use a reference clip up to 10.33 seconds long, a scene-specific prompt and a seed to explore controlled video transformations.
Turn a still image into a cinemagraph with selective motion and a fixed camera. Generate up to 121 frames at 24 fps for product loops, ambient visuals and social posts.
Reduce macroblocking, color bleed and compression artifacts. Transform short reference clips with scene-specific prompts and a seed for reproducible tests and production previews.
Bring defocused footage back into sharper focus. Use a reference clip up to 5.33 seconds long, a scene-specific prompt and a seed to explore controlled video transformations.
Relight a daytime shot as a night scene. Use a reference clip up to 4.33 seconds long, a scene-specific prompt and a seed to explore controlled video transformations.
Add plausible natural color to black-and-white footage. Transform short reference clips with scene-specific prompts and a seed for reproducible tests and production previews.
Remove people from footage to create an empty scene. Use a reference clip up to 5.33 seconds long, a scene-specific prompt and a seed to explore controlled video transformations.
Expand videos into 7 aspect ratios with generated surroundings while keeping source timing and audio. Guide the added scene with a prompt and export at up to 1080p.
Pin up to 10 images to exact frames and get one continuous 5 to 20 second video at up to 1080p that passes through each of them, with native sound and music.
Animate a still image into a 5 to 20 second video at up to 1080p with native sound and dialogue, or add a Last Frame to control exactly where the shot ends.
Generate 5 to 20 second cinematic videos at up to 1080p from a prompt, with native dialogue in many languages, sound effects and music synced to the action.
Generate 4 to 15 second videos with native dialogue and sound at 480P or 768P, from a prompt, a first and last frame, or reference images, videos and audio.
Animate a person, pet or character from a single image so they speak or sing to your audio track, with lip sync, gestures and camera moves you direct in the prompt.
Edit or combine up to 8 images from plain instructions: recolor, relight, swap text, remove objects or restyle, while everything you did not mention stays untouched.
Generate clean 3D meshes from text or one image at the face count you set, from a few hundred to 30,000 faces, with 2K to 8K textures and no remesh step.
Turn a video of up to about 6 seconds into a grayscale alpha matte that separates the subject from the background, ready to use as a luma matte for compositing.
Add synced sound effects, foley and ambience to a video and get it back with sound. Keeps existing dialogue, replaces other audio. Clips up to 10 min, up to 4 variations per run.
Generate sound effects, foley and ambience synced to your video, returned as an audio track. Keeps existing dialogue on request. Clips up to 10 min, up to 4 variations per run.
Generate sound effects and ambiences from a text prompt, from quick one-shots to 10-minute beds. Up to 4 variations per run, negative prompt control, WAV, MP3, AAC or FLAC.
Finalize a Seedance 2.5 480p draft into a 1080p MOV with the same motion, framing and sound, without regenerating. Iterate on cheap drafts and upscale only the keepers.
Turn Blender renders and 3D previs into photorealistic video: your clip sets the layout, camera, motion and cuts, while a prompt and up to 8 references set the look, up to 1080P.
Turn one portrait into a lip-synced talking video up to 3 minutes long. Write a script for 30 voices in 10 languages, or upload your own speech or song as the audio.
Generate 5 to 15 second videos guided by up to 9 images, 3 clips and 3 audio files. Keep products and characters exact, copy a clip's motion, or keep your own soundtrack.
Turn ideas and up to 5 reference images into prompts and parameter proposals for a target model. Explore alternatives with fast or thorough processing, then review and run.
Convert video luminance into a grayscale mask clip. Keep continuous tones, set a threshold from 0 to 255, or invert light and dark regions for motion graphics.
Convert image alpha or luminance into a grayscale mask. Preserve soft transitions, apply a threshold from 0 to 255, or invert the result for your compositing workflow.
Redo part of a clip with Lightricks' LTX-2.3 22B: pick a time range and regenerate its video, audio or both, like a new line of dialogue, keeping the rest intact.
Fast distilled tier of Lightricks' LTX-2.3 22B: quick text or image to video with synced audio and end frame support. Up to 20 s at 1080p, ideal for drafts.
Turn a voice or music track into video with Lightricks' LTX-2.3 22B: lip-synced talking characters or performers that follow the rhythm, from an optional image.
Lightricks' open-weights 22B video model: text or image to video with synced speech and sound, start and end frame control, and a negative prompt. Up to 4K.
Tripo Smart UV re-unwraps a 3D model into a clean, organized UV layout, bakes its textures onto it and keeps the rig intact, with a UV atlas preview on every result.