Try the tools right here
Each panel below is a live miniature of the real thing — drag, click and slide it.
Auto Coloring
Upload line art, and AI colors it in one pass.
Rough coloring eats the first hours of every concept. MergeRef detects hair, skin and outfit regions in your line art and fills them naturally — with an intensity slider that runs from fully automatic to lightly guided.
Five generation steps, each with its own result image slot — drop the real renders in and the after side follows the slider.
Tools built for production, not demos
Pose Extraction
Detect joint structure from any reference, extract the skeleton, and put your own character in the same pose.
Outfit Extraction
Lift armor, dress or uniform detail from a reference and composite it onto a different body type or pose.
Masking & Map Conversion
Depth maps, OpenPose skeletons, character masks and PBR maps — computed locally with on-device AI.
Video Generation
Turn canvas concepts into motion with Veo 3.1 and Seedance 2.0, straight from your references and maps.
Object Insert · 7
Drop props, weapons or set pieces into a finished illustration. The model composites them in with matching perspective and light instead of pasting them flat.
Inpainting · 8
Brush a mask over anything — a hand, a helmet, a background patch — and regenerate only the painted area. Use it when a body region isn't the shape you need.
Photoshop-style panels
Tear history, source image, workflow and inspector panels out into their own OS windows, snap them together, and dock them back when you want the single-window layout.
Workflow & prompt libraries
Save prompt, role images and options as a reusable workflow; keep quality and style phrases as prompt styles. A1111 syntax — (emphasis:1.2), [de-emphasis], <lora:name> — works as typed.
Connect your own ComfyUI
ComfyUI isn't bundled — if it's already installed on your PC, MergeRef detects it, launches it, and generates with your own SD 1.5 / SDXL checkpoints at zero API cost, with an inspector for checkpoint, sampler, seed, LoRA and ControlNet.
Local mode consumes the original image only — REF, outfit, pose and style reference roles are ignored, and pose comes from the pose tool's control map instead.
Community node workflows
Search and load community templates as a node graph, assign canvas images to node inputs, and let your ComfyUI-Manager download any missing models or custom nodes from inside MergeRef. Requires your own ComfyUI install.
One payment. Yours forever.
$19.99 once — every tool on this page included, with your own API keys.