Segment videos using Meta's SAM 3. Supports prompt-based and tracking-based segmentation.
Models
All Models
Segment images using Meta's SAM (Segment Anything Model) 3. Supports prompt-based and tracking-based segmentation.
Wan 2.6 T2V by Alibaba is a commercial model for 1080p multi-shot video with native audio and lip-sync. Generation from 75 CU.
Wan 2.6 I2V by Alibaba is a commercial model that animates images into 1080p videos with native audio and lip-sync. Generation from 75 CU.
Magnific Upscaler Precision is an AI tool that increases resolution while adding high-fidelity detail
Magnific Creative is an interpretive/creative upscaler that synthesizes new textures and high-end detail from 16 credits per image.
Photoroom Uncrop reconstructs clipped subjects or missing object parts that were cut off at the image edge.
Photoroom Text Remover erases text overlays and logos from images while reconstructing the underlying background.
Photoroom Relighting simulates new light sources to adjust exposure and mood while preserving brand color accuracy.
Photoroom Expand extends an image’s background to fill a new canvas size without stretching the original content.
Photoroom Background Replacer generates environments for subjects using text prompts or reference images.
Photoroom Background Removal creates precise subject cutouts by automatically identifying the primary object in a photo.
Kling O1 Reference Video by Kuaishou uses a source video for style/motion and can combine it with images or Elements to generate new, consistent scenes.
Kling O1 Video Editing by Kuaishou modifies video clips using text prompts and can incorporate images or Elements for fine-grained content manipulation.
Kling O1 Reference Images by Kuaishou generates video using multiple images to define a character or object, ensuring consistency. Pricing from 90 CU.
Creatify Aurora by Creatify is an image-to-video model for ultra-realistic reactive avatars with emotive gestures. Currently unavailable.
Minimax Music 2.0 creates full songs with vocals and instruments from text, generating up to 5-minute tracks. Pricing starts from 5 CUs.
Sync Lipsync React-1 by Sync Labs provides audio-driven lip-sync with integrated emotion presets and head movement control.
FLUX 1.1 (Pro) by Black Forest Labs has enhanced performance with improved speed and detail over the original Flux 1 series. Pricing from 6 CU.
FLUX 1.1 (Pro Ultra) by Black Forest Labs does 4MP resolution with a specialized "Raw" mode for candid, authentic photography styles. Pricing from 9 CU.
Kling AI Avatar 2 (Pro) by Kuaishou generates a lifelike talking video from a single static image and an audio track, focusing on precise lip-sync.
Kling 2.6 I2V (Pro) by Kuaishou animates images into 1080p video with synchronized native audio and advanced motion. Pricing from 105 CU.
Kling 2.6 T2V (Pro) by Kuaishou generates 1080p video with synchronized native audio directly from text prompts. Pricing from 105 CU.
Kling O1 I2V by Kuaishou creates 1080p video from images with precise start/end frame control for structured animation. Pricing from 90 CU.
This Scenario model creates voxel-based 3D models from prompts with simple controls for size and scale.
Grid Maker by Scenario organizes multiple images into a clean, adjustable grid layout for a unified and clear visual overview.
Scenario Video to Image Sequence converts video footage into high-quality, frame-by-frame image sequences with customizable intervals for seamless asset generation.
Sequence-to-Video by Scenario turns a collection of images into a video file with settings to control the playback speed, looping style, and sound.
Scenario Gemini Upscale by Google uses multimodal reasoning to sharpen images while balancing structure and creativity from 36 credits.
This Flux LoRA generates sci-fi landscapes using bold lines and cartoon aesthetics. It produces any landscape or environment with smooth shading.
This LoRA generates glossy, paint-like surreal scenes with dripping liquids, swirling textures, vibrant florals, and candy-colored spheres in dynamic abstract spaces.
This Flux LoRA generates stylized 3D cartoon characters like robots and wizards. It uses vibrant colors and exaggerated proportions for games and animation.
This Flux LoRA produces semi-realistic costume collections. It generates any historical or professional gear from helmets to boots in muted, natural tones.
Z-Image Turbo (Dec 2025) is a distilled 1-step model for nearly instant, real-time image generation with higher resolution than older LCMs. 2 CU only
LTX-2 Retake by Lightricks enables precision re-rendering of video segments to fix performance or dialogue without full regeneration.
Image Slicer by Scenario cuts an image into a custom grid of smaller sections and saves each one as its own separate file.
Pixel Snapper cleans up pixel art by fixing scaling issues and keeping the blocks aligned for a sharp, retro look.
A specialized model built for generating seamless, top-down tilemaps with perfect grid alignment and terrain coherence.
A premium pixel-art generation model designed for exceptional quality, rich detail, and stylistic accuracy.
A dedicated animation model featuring specialized motion-aware generation, perfect for sprite sequences.
FLUX 2 (Dev) by Black Forest Labs is an open-weight model balanced for experimentation and custom training workflows. Pricing from 3 CU.
Meshy Rigging generates skeletal structures and skin weights for 3D meshes to enable character animation.
Hunyuan 3D Part is a 3D-to-3D segmentation model that breaks existing meshes into organized, editable sub-parts.
Gemini 3.0 Pro Edit by Google offers professional-grade conversational editing with deep context awareness starting from 22 credits.
Flash VSR Upscale Video by Tsinghua/Shanghai AI Lab provides real-time diffusion-based upscaling with a one-step streaming framework.