Generate a 3D model from a photo, sketch, or concept image. Choose Studio, the visual Canvas, an MCP-connected agent, or the CLI.
Example preview · Drag to rotate · Use + / − to zoom
Compare each reference image with the generated 3D model. Open a result and rotate it to inspect the silhouette and the surfaces AI inferred beyond the original view.
Use Studio for a guided browser workflow: upload a reference, choose settings, and inspect the generated model.
Choose a photo, sketch, or concept image with one complete subject. Consistent front, back, left, and right views can use a compatible multi-view option.
Select a generation model, review its options and credit cost, then submit the task. Studio shows progress while it runs.
Rotate the model to check its silhouette and inferred surfaces. With an eligible paid account, download an available format or continue refining it.
Choose the interface that fits your image-to-3D task. Use Studio for browser controls, Canvas for connected visual steps, MCP for an AI agent, or CLI for a terminal or script.
Read-only Character Design template · generated sample assets
VISUAL WORKFLOW
Keep the reference, model, and material steps together on the Canvas. Connect an image to Image to 3D, pass the model to Texture, and run each node as needed.
Use Kreat3D to generate a textured 3D model from this reference image. Keep the main silhouette and visible details, then show me the available files when the task is complete.
AGENT AND TERMINAL
MCP lets an AI agent call Kreat3D tools, follow the task, and retrieve available output files. Use the CLI from a terminal or script.
Choose how your image becomes a 3D model. Start with one reference or consistent multi-view images, then select geometry, texture, and export options for the next step.
Start with a photo, sketch, or concept image that shows one complete subject. The AI uses visible shape and detail to build the model and infers surfaces the image does not show.
Add consistent views of the same object to give the generator more visual context. Multi-view support depends on the selected model and an eligible subscription.
Generate geometry only or choose a supported textured option. Texture quality, PBR materials, and geometry controls vary by model.
Continue in Blender, a game engine, or a 3D printing workflow with an available format. Download access requires a paid account, and format availability varies by model.
Upload, configure, and inspect the result in your browser without installing a modeling app. Studio shows task progress while generation runs in the cloud.
Use an image-generated 3D model as a starting point for game prototypes, product visualization, 3D printing, AR, animation, or teaching.
Turn character art and prop concepts into assets for prototyping. Review topology, optimize the mesh, and add rigging where your game needs it.
Explore form and proportions from product references before the next review. Use the output for visualization, then develop precise dimensions in your design tools.
Start a figurine or object from a clear image. Check scale, wall thickness, and mesh integrity in your preparation tools before slicing.
Explore interactive product presentation with 3D assets. Check material accuracy and optimize file size for your viewer or AR platform.
Move a visual concept into scene development. Retopologize, rig, and refine materials in your animation package when the shot requires it.
Turn visual references into objects learners can explore from different angles. Review accuracy before using models to explain technical subjects.
Continue your workflow in
Prepare a clear reference before you generate a 3D model. A complete subject, readable depth, even lighting, and consistent extra views give the AI more useful visual information.
Keep the object in frame against a simple background. Avoid cropping defining parts such as handles or limbs.
Shading can communicate volume better than a flat outline. If you only have an idea, create a reference first.
Keep edges visible. Deep shadows, glare, blur, and occlusion leave more of the shape open to interpretation.
Across multiple images, keep the object, materials, scale, and design consistent.
No reference image yet? Create concept art before you generate the model.
Create a Reference ImageFind answers about 2D image inputs, multi-view generation, textures, Canvas and agent workflows, free starter credits, and model exports.
It creates 3D geometry and, with a supported option, textures from a 2D reference image. Because one image does not show every surface, the AI infers hidden areas and the result may need refinement.
Yes. Use a clear photo, JPG, PNG, WebP, sketch, or concept image with one complete subject. A simple background, visible edges, even lighting, and readable depth give the generator more useful visual information.
Yes. Multiple Images combines consistent views of one object. The front is required; back, left, and right views are optional. It requires a compatible model and eligible subscription.
The editor shows an estimate before submission. Actual time varies with the model, settings, object complexity, and queue. Studio displays progress after the task starts.
Choose a supported textured option when you want surface color with the mesh. Geometry-only output is untextured; texture quality and PBR options vary by model.
Use Studio for browser controls, progress, and inspection. Use Canvas for connected image, model, and texture steps; MCP for a compatible AI agent; or CLI for a terminal or script.
Yes. Connect an image to Image to 3D, or select an existing account model from My Assets. Add Texture when needed, then run or re-run the relevant nodes.
Yes. Through MCP, a connected AI agent can create the model, follow progress, and return available output. Use the Kreat3D CLI for a terminal workflow; the MCP and CLI page covers setup.
New accounts receive starter credits to try supported generation options without entering card details. Each task's credit cost appears in the editor. Starter credits do not unlock every model, multi-view mode, or generated-model downloads.
Export paths include GLB, OBJ, FBX, and STL where supported; availability varies by model. After import, check scale, topology, and materials for Blender or game use. For 3D printing, also check dimensions, wall thickness, closed surfaces, orientation, and supports.
Upload a photo, sketch, or concept image and generate a 3D model you can inspect, refine, and carry into your next workflow.
Free starter credits · No card required to try · Generated-model downloads require a paid account.
