Sync-3 Lipsync by Sync Labs syncs any audio to a speaker's mouth in video. Built for dubbing, voice-over, and ADR with five duration-mismatch modes.
Models
All Models
JoyAI Image Edit by JD Open Source: edit any photo with a plain-language instruction. Control guidance, inference steps, and negative prompts for precise results.
P-Image Upscale by Pruna AI. Fast upscaling up to 8 MP via target megapixel or side-factor mode, with optional detail and realism enhancement.
Wan 2.7 Video Edit by Alibaba rewrites video content from text instructions: swap backgrounds, shift lighting, apply styles, or restyle using a reference image.
Wan 2.7 Image Pro by Alibaba generates up to 4K images from text and edits or fuses up to 9 reference images, with optional thinking mode for deeper prompt reasoning.
Wan 2.7 Image by Alibaba: text-to-image and multi-reference editing at 1K or 2K, with image set mode and thinking mode for up to four outputs per run.
Minimax Speech 2.8 Turbo by MiniMax. Fast TTS with 17 voices, 10 emotions, and 40+ language boosts, built for low-latency apps.
Minimax Speech 2.8 HD by MiniMax. Premium TTS with 17 voices, 10 emotions, 41 languages and dialects, natural interjections, and precise control over speed, pitch, and volume.
Hy Wu Edit by Tencent. Transfer outfits, swap faces, and blend textures using up to 3 reference photos, with no fine-tuning required.
Google Lyria 3 Pro generates complete, structured songs up to 2 minutes from text prompts or up to 10 reference images, with control over genre, tempo, vocals, and sections.
Google Lyria 3 Clip by Google generates 30-second music clips from text prompts or reference images, with control over genre, tempo, mood, and song structure.
Grok Edit Video by xAI: edit any clip with a text prompt. Restyle scenes, swap objects, change environments, all while preserving what you don't touch.
Grok Extend Video by xAI continues your clip from its last frame, generating 2 to 10 seconds of new AI footage guided by a text prompt.
Flux.2 LoRA for stylized 3D costume sets. Renders outfit collections and equipment on invisible figures with detailed textures and studio lighting.
Tada 3B by Hume AI clones any voice from a short audio reference and synthesizes multilingual speech across 10 languages with no transcript hallucinations.
Tada 1B by Academia / Open Source clones any voice from a short audio sample and synthesizes multilingual speech across 10 languages with speed control.
HeyGen Avatar 4 by HeyGen: turn a face photo into a talking video with 80+ preset voices, resolutions up to 1080p, and stable or expressive motion.
HeyGen Video Translate Speed dubs your video into 170+ languages with AI lip sync, optimized for fast, high-volume translation at scale.
HeyGen Video Translate Precision by HeyGen. Translate video speech into 170+ languages with high-fidelity lip sync, voice cloning, and multi-speaker support.
HeyGen Video Agent by HeyGen turns a text prompt into a polished presenter video, handling scripting, avatar selection, scenes, and narration pacing automatically.
Pixelcut Background Removal by Pixelcut: isolate subjects from any image with clean edge detection. Export as full RGBA composite or alpha-only mask.
Physic Edit applies physics-aware transformations to images: flood cities, melt armor, shatter glass, or freeze scenes with accurate refraction and deformation.
A Flux 2 LoRA for 3D low-poly environments suited to platformers and adventure games, with geometric shapes, vibrant colors, and soft lighting.
Flux.2 LoRA for stylized game icons: gemstones, crystals, and enchanted loot rendered with vibrant colors and polished digital painting.
Generate editable SVG glyphs that match your brand's typography. Upload up to 8 reference characters, type your target, get a clean vector output up to 4K.
VecGlypher generates editable SVG glyphs from a text style description. Control fill, stroke color and width, and output size up to 4096 px.
LTX 2.3 Pro Retake by Lightricks. Regenerate any section of a video: replace audio, video content, or both. Define the start time and duration, prompt the change. 1080p.
LTX 2.3 Pro Extend Video by Lightricks adds new footage to the start or end of any clip. Guide the continuation with a prompt and choose 6, 8, or 10 seconds.
Trellis 2 by Microsoft retextures any 3D mesh using a reference image, producing PBR-quality textures at up to 4K with adjustable guidance and resolution.
Kling V3 Pro by Kuaishou. Transfer motion from any reference video onto a character image, with facial consistency and optional audio preservation.
Kling V3 Standard Motion Control by Kuaishou. Animate any character image using a reference video to drive their movements, with optional audio transfer.
Tripo 3.1 Multi View by Tripo AI generates accurate, PBR-ready 3D meshes from up to four reference angles: front, left, back, and right.
Tripo Rigging 2.5 by Tripo AI auto-rigs non-biped creatures (quadruped, hexapod, octopod, avian, serpentine, aquatic) and optionally retargets with preset animations.
Vidu Q2 Pro by Shengshu Technology generates videos from up to 7 reference images or 2 reference videos, preserving subject identity at up to 1080p.
Tencent UV Unwrapping by Tencent automatically generates clean UV maps for FBX, OBJ, and GLB models up to 30,000 faces, ready for texturing.
Tencent Texture Edit by Tencent. Retexture FBX models using a text prompt or reference image. Prompt mode outputs full PBR maps for game-ready assets.
Qwen Edit 2511 by Alibaba edits images from plain text instructions. Add, remove, or restyle elements, with up to 6 LoRA styles and multi-reference image support.
FLUX.2 Klein 9b by Black Forest Labs. Fast text-to-image and image editing with strong style coherence, multi-reference support, and LoRA compatibility.
Vidu Q3 Turbo by Shengshu Technology animates images into 1080p video up to 16 seconds, with start/end frame control and optional background audio.
Vidu Q3 Pro by Shengshu Technology animates a single image or interpolates between a start and end frame, up to 16 seconds at 1080p.
Vidu Q3 Turbo T2V by Shengshu Technology. Fast text-to-video generation up to 16 seconds and 1080p, with optional background music and flexible aspect ratios.
Vidu Q3 Pro by Shengshu Technology. Premium text-to-video up to 16 seconds in up to 1080p, with cinematic camera control and optional audio.
Seedream 5.0 Lite by Bytedance. Fast text-to-image and image editing at 2K, 3K, or 4K. Strong instruction following, multi-reference inputs, and sequence generation.
Runway Gen4.5 by Runway ML: cinematic text/image-to-video with precise motion control and delivery up to 12-bit HDR ProRes or PNG sequence.
Tripo Stylization by Tripo AI converts any existing 3D model into LEGO, Voxel, or Voronoi style in one step.
Tripo Retopology by Tripo AI converts dense or AI-generated 3D meshes into clean, low-poly geometry with face count control and texture baking.
Tripo 3.0 Texturing by Tripo AI applies PBR materials or painted textures to existing 3D models using a text prompt, image, or style reference.