Trending

Enhance muffled 16 kHz speech audio into crystal-clear 48 kHz, with denoising for particularly bad inputs.

Image generation with BitDance. Fast, high-resolution photorealistic images using an autoregressive LLM— for efficient, high-quality results.

PersonaPlex is a real-time, full-duplex speech-to-speech conversational model that enables persona control through text-based role prompts and audio-based voice conditioning.

FireRed Image Edit is FireRed's state of the art open source editing model, re-trained from Qwen Image Edit 2509.

wan-motion

Wan Motion is a streamlined character animation model that transfers motion from a driving video onto a reference character image. Based on Wan-Animate which preserves the original character's proportions, Simple uses pose retargeting to adapt the driving video's skeleton to match the reference character's body shape, producing more natural results when the two have different builds. It outputs at 720p with optimized defaults for fast, high-quality generation — just provide a video, an image, and an optional prompt.

GenFocus Model to Refocus Images

GenFocus Model to Refocus Images

PersonaPlex is a real-time, full-duplex speech-to-speech conversational model that enables persona control through text-based role prompts and audio-based voice conditioning.

qwen-image-trainer-v2

Qwen Image LoRA training

z-image-base-trainer

Fast LoRA trainer for Z-Image, a super fast text-to-image model of 6B parameters developed by Tongyi-MAI.

qwen-image-max/text-to-image

Text-to-Image endpoint for Qwen-Image-Max. Qwen Image Max improves upon the Qwen Image Plus series by enhancing the realism and naturalness of images.

qwen-image-max/edit

Image editing endpoint for Qwen-Image-Max. Qwen Image Max improves upon the Qwen Image Plus series by enhancing the realism and naturalness of images.

flux-2-klein-4b-base-trainer

Fine-tune FLUX.2 [klein] 4B from Black Forest Labs with custom datasets. Create specialized LoRA adaptations for specific styles and domains.

flux-2-klein-4b-base-trainer/edit

Fine-tune FLUX.2 [klein] 4B from Black Forest Labs with custom datasets. Create specialized LoRA adaptations for specific editing tasks.

qwen-image-2512-trainer-v2

Fast LoRA trainer for Qwen-Image-2512

Enhance muffled 16 kHz speech audio into crystal-clear 48 kHz

LongCat-Video-Avatar is an audio-driven video generation model that can generates super-realistic, lip-synchronized long video generation with natural dynamics and consistent identity.

UltraShape-1.0 is a 3D diffusion framework that generates high-fidelity 3D geometry through coarse-to-fine geometric refinement.

ltx2-video-trainer

Train LTX-2 for custom styles and effects.

LongCat-Video-Avatar is an audio-driven video generation model that can generates super-realistic, lip-synchronized long video generation with natural dynamics and consistent identity.

Semantic image alignment measurements

Image reference comparison measurements

Reference-free image measurements

wan-move

Use Wan-Move to generate videos with controlled the motion using trajectories