Two images in, one out. Pick style transfer or object fusion. Drop a subject and a reference, pick your mode. Style locks the subject. Object blends them into one.

Generate 8 design variations from one image in a single shot. Drop an image, pick an axis (colors, proportions, shape, or style), generate. Plnty hack on GPT Image 2. Great for exploring directions fast.

Recolor any image to match a custom palette of your choice. Set the colors, drag the weights to tune dominance, generate. Great for brand exploration or quick color variations.

Combine two images, a subject and a scene. Drop the subject onto the scene and the AI harmonizes light, perspective, and shadow. Rotate, scale, and position the object in your scene however you like.

Generate 4 curated close-ups from one image in a single shot. Drop an image. Get back construction, material, edge, and functional details. Pick LQ (512²) or HQ (1024²) on the spawned card.

Drop any image to get back a porcelain twin of it. The tool describes your image and recreates it as a softer, dreamlike version. Works best on portraits, still life, and single objects.

Fast, high-quality base text-to-image from Black Forest Labs.

Flagship text-to-image model with industry-leading quality. Write a prompt. Get back high-fidelity images with full format control. Black Forest Labs' top model. Pick this for hero outputs.

Maximum performance: Highest editing consistency across tasks. Strongest prompt following and faithful representation of various styles. Supports grounded generation with real-time web context.

Design-taste text-to-image with strong prompt accuracy. Write a prompt. Get back clean compositions with good text rendering. Best for posters, layouts, and graphics that need readable type.

Lightweight, fast text-to-image from the Seedream family. A quick, economical pick for drafting and iterating on ideas before you commit to a heavier model.

Alibaba's Qwen Image foundation model with strong text rendering. Write a prompt. Get back images with clean typography baked in. Best when readable text inside the image is the priority.

OpenAI's GPT Image 2, a strong prompt-following text-to-image model that's reliable with on-image text. Best for detailed, instruction-heavy prompts where the model needs to follow exactly what you asked.

Google's Nano Banana, a quick and reliable everyday text-to-image model. Dependable for fast, clean results when you don't need the heavier options.

Krea 2 with style-reference steering. Drop up to 10 reference images and the model blends their look into new generations. Best when you want output that matches a specific style or mood board.

Plnty's house text-to-image style model, a distinctive in-house aesthetic tuned by the team. Reach for it when you want our signature look instead of a generic model style.

Edit images with Nano Banana 2, the improved successor. Drop your images, write what you want changed, generate. Sharper edits than the original Nano Banana, with stronger fidelity.

Premium image editing with Nano Banana Pro (Gemini 3 Pro Image). Make precise, position-faithful edits that respect your original composition. Swap elements, adjust details, and refine without the layout drifting.

Edit images with OpenAI's GPT Image 2. Up to 4 reference images and an optional mask. Generate. Best when you need precise mask-driven local edits.

Edit a single image with Flux Kontext. Drop one image, write a prompt, generate. Specialized in context-aware edits and consistent style preservation.

Generate or edit images with xAI's Grok Imagine. Drop your image, write a prompt, generate. Distinct visual character. Good for experimental work.

Generate a 3D mesh from a single image with Tripo 3.1. Drop one image. Get back a quad-topology mesh with rich detail. Best when you need clean topology for editing or animation.

Generate a game-ready 3D mesh in around two seconds. Drop one image. Get back a low-poly mesh fast. Use it for prototyping or when speed matters more than detail.

Generate a 3D mesh from a text prompt with Tripo. Describe what you want. Get back a clean quad mesh with PBR. Best when you need editable topology or smart quads.

Generate a high-quality 3D mesh from one image with Meshy V6. Drop one image. Get back a mesh with full PBR textures. Best when you need photoreal materials and detail.

Generate a 3D mesh from 1 to 4 reference images with Meshy V6. Drop your references from any angles. The tool figures out the geometry. More flexible than orthographic multiview. Good for partial references.

Generate a 3D mesh from a single image with Hunyuan 3.1. Drop one image. Get back a textured mesh with PBR maps. Fastest on Replicate. Good first pick for most subjects.

Generate a draft 3D mesh from a single image with Trellis. Drop one image. Get back a quick mesh, light on detail. Use it for fast sketches and iteration. Switch up for final.

Generate a 3D mesh from a single image with Trellis 2. Drop one image. Get back a fast mesh that handles tricky subjects. Great for unusual shapes and stylized objects.

Generate a 3D mesh from a single image with Rodin Gen-2.5. Drop one image. Get back a mesh with detailed shading and PBR. Strong on hard surface, manufactured, and industrial objects.

Reconstruct a textured 3D model from a turntable video. Drop a 360-degree video clip. The tool builds the mesh and textures. Best when you have clean lighting and a controlled turntable.

Generate multiple 3D objects from a single scene image. Drop a scene image. The tool segments and generates each object separately. Great for breaking a single render into independent 3D pieces.

Rebuild a 3D mesh with cleaner, more even topology using Meshy. Turn dense or messy geometry into a tidy layout that's far easier to edit, texture, and animate.

Industry-leading AI textures with full PBR output from Meshy. Drop a mesh, write a prompt, get a full PBR retexture. Best when you need photoreal materials on hero models.

Repaint a 3D model with Tripo using prompt, image, or multi-view. Drop a mesh and your references. Choose your control mode. Generate. Most flexible retexture tool. Great when you have lots of references.

Image-guided texturing up to 4K resolution with Trellis 2. Drop a mesh and a reference image. Generate a high-res texture. Best for sharp, photoreal surface detail at print resolution.

Generate PBR maps from a photo as canvas image references. Drop an image. Pick Patina or PBR Estimate. Generate maps. Output lands as image refs you can wire into any 3D node.

Restyle an existing 3D mesh into a new artistic look with Tripo. The underlying shape stays put while its surface and detail take on a different aesthetic. Drop a model, pick a direction, regenerate.

Generate a video from an image with Kling o3. Drop an image, write a motion prompt, generate. Reliable workhorse model. Good first pick for most clips.

Generate a video from an image with Kling o3 Pro. Drop an image, write a motion prompt, generate. Higher quality than standard Kling. Pick this for hero shots.

Generate a video from an image with Seedance 2.0 Fast. Drop an image, write a motion prompt, generate. Fast, snappy clips. Great for quick experiments.

Chat with any major LLM: Claude, GPT, Gemini, or Llama. Write your prompt. Optionally add an image for multimodal queries. Pick the model that fits the task. Switch anytime on the card.

Rewrite a rough prompt into something more detailed. Drop your draft prompt. Get back a polished version for image or 3D gen. Use it before any text-to-image or text-to-3D generation.

Get curated reference sites that match your description. Write what you're looking for. Get back relevant URLs. Faster than Pinterest. Great for early-stage inspiration.