3D Studio is Sorceress’s workspace for creating smooth polygon 3D mesh models from images or prompts, then preparing those models for game and animation use. If you want to “turn an image into a 3D model,” start here: the Create tab can generate textured mesh assets that you can inspect, optimize, rig, refine, animate, and download in common 3D formats when available.
Looking for blocky cube/voxel models instead? 3D Studio creates polygon mesh models. For voxel-style cube models, use Voxel Studio instead.
Turn an image into a 3D model
The fastest workflow begins in the Create tab.
- Open 3D Studio.
- Go to Create.
- Choose an Input Mode in the right panel:
- Image — generate a model from one image.
- Multi-Img — generate from several reference views of the same subject.
- Text — generate from a written prompt.
- Select one or more generation models in the left Select Models panel. Only models compatible with the current input mode can be selected.
- Provide your input:
- For Image, drag image files into the upload area or click it to choose files. Dropping multiple images stages them as separate generation jobs.
- For Multi-Img, upload 2–4 views. Some models show explicit Front, Left, Back, and Right slots; the Front slot is required in that layout.
- For Text, type a prompt describing the object or character you want.
- Adjust any model settings you want to use.
- Click Generate or Start Job.
- When the model appears in your gallery, click its card to open the viewer.
- Use the model tools to inspect, download, optimize, or send it to Rig.
For characters you plan to rig, use a clean front-facing image whenever possible. A neutral standing pose with visible arms and legs gives the rigging tools the best chance of building a usable skeleton and weights. The Create panel includes a recommended A-pose reference card and a downloadable pose reference image to help you prepare source art.
Input modes
Image
Use Image mode when you have one clear reference image. This is the most direct way to convert an image into a textured 3D mesh.
You can drop multiple images in Image mode. Each staged image becomes its own generation job, so this is useful when you want to test several characters or props with the same model settings.
Each staged image can have its own label. You can also add a batch label so the resulting gallery cards are easier to identify later.
Multi-Img
Use Multi-Img when you have several views of the same object or character. Depending on the selected generation model, the interface either accepts an unordered set of 2–4 images or shows view slots such as Front, Left, Back, and Right.
Tips for multi-image input:
- Use consistent lighting and scale across all views.
- Keep the subject centered and unobstructed.
- Do not mix different outfits, poses, or versions of the subject.
- If labeled slots are shown, provide the Front view and add side/back views when available.
Text
Use Text mode when you do not have reference art. Type a concise prompt describing the 3D object, creature, character, or prop. The prompt field supports up to 600 characters.
Text mode is best for conceptual props and exploratory assets. For a specific character design you already have, image-to-3D usually provides a more faithful result.
Choosing generation models and settings
The left panel lists the available generation models. Some models support all input modes; others only support image or text. If a model is incompatible with the current input mode, it appears unavailable until you switch modes.
Many models include a settings button. Settings vary by model, but visible controls may include:
- Topology / Mesh Mode — choose whether the result should prioritize standard triangle mesh output or cleaner quad-like topology when supported.
- Target Polycount / Face Limit / Mesh Density — control how dense the generated mesh should be. Higher density can preserve detail but creates heavier files. You can optimize later in the model viewer.
- Symmetry — allow automatic symmetry detection, force symmetry, or disable it when supported.
- Pose Mode / Force T/A-Pose — ask the generator to place a humanoid character in a more rig-friendly pose.
- Mesh Type — choose a standard detailed mesh or a low-poly style when supported.
- Texture — choose whether to generate textures and, where available, the texture quality.
- PBR Maps — generate physically based material maps for more realistic lighting in engines that support them.
- Remesh — rebuild topology for cleaner geometry when the selected model supports it.
- Image Enhancement — improve the input image before reconstruction.
- Remove Lighting / Remove Background — reduce baked-in lighting or remove the source background before reconstruction.
- Texture Alignment / Orientation / Auto Size — control how textures align and how the output is oriented or scaled when supported.
- Seed / Randomize Seed — reproduce or vary results for models that expose seed controls.
- Quality presets — some models use simple Low / Medium / High style presets instead of individual sliders.
- Public / Private visibility — when shown, this controls whether a generated model is eligible to appear in shared asset browsing areas.
You do not need to understand every setting to begin. A practical starting point is: choose a recommended or default model, use a clear source image, keep texture generation enabled when you want a finished-looking model, and optimize after generation if the mesh is too heavy.
Create gallery
Generated models appear as cards in the Create gallery. The gallery includes:
- All / Rigged / Unrigged filters — quickly find models that have or have not been rigged.
- Search — filter by generation label.
- Grid size control — adjust how many cards appear per row on larger screens.
- Before/after thumbnails — when both the source image and rendered thumbnail are available, drag the thumbnail divider to compare the input image against the generated 3D result.
- Rename — click a card label to rename it.
- Delete — remove a generation from your gallery.
- Retry — retry a failed generation when available.
- Reuse source — place a previous source image back into the upload area for another run.
- Rigged badge — shows when a generated model already has a saved rigged character associated with it.
Click a completed generation card to open its detail viewer.
Inspecting a generated model
The model detail viewer lets you review the generated mesh before rigging or downloading it.
Visible viewer tools include:
- View mode toolbar — switch how the model is displayed, such as textured or alternate inspection views when available.
- Rotate slider — rotate the preview without changing the underlying asset.
- Source comparison blind — when a source image is available, turn on Source and drag the slider to overlay the original image over the 3D model.
- Original / Optimized / Both — when an optimized version exists, compare it side by side with the original.
- Loading meter — large models show download and preparation progress while loading into the viewer.
- Model statistics — the inspector panel shows geometry information such as triangle counts when available.
Use the viewer to check whether the silhouette, proportions, texture placement, and polygon density are suitable before investing time in rigging.
Optimizing and downloading models
The right-side inspector panel for a completed model includes download and optimization tools when the model has loaded successfully.
Depending on the generated output, downloads may include formats such as GLB, FBX, OBJ, USDZ, or other model files. The viewer detects the actual available format and labels downloads accordingly.
Optimization tools let you create a lighter version of the mesh for real-time use. After optimization, the viewer can show:
- Original — the unoptimized generated model.
- Optimized — the reduced version.
- Both — side-by-side comparison.
Compare the optimized mesh carefully. A good optimized model should preserve the important silhouette and texture appearance while reducing unnecessary geometry.
Uploads
The Uploads tab is for models you bring into 3D Studio yourself. Uploaded models follow the same current-character workflow as generated models:
- Open Uploads.
- Choose an uploaded model.
- Send it to Rig if it is unrigged.
- Save the rigged result.
- Open it in Refine or Animate once it has a rig.
If an uploaded source already has a saved rigged version, the studio can open the rigged character directly instead of making you rig the original again.
Rig
The Rig tab turns an unrigged mesh into an animation-ready character by adding skeleton and skin-weight data. Rig becomes available after a model is open from Create, Uploads, or another supported entry point.
Use Rig when you want to:
- Build a skeleton for a generated or uploaded mesh.
- Edit an existing saved rig.
- Recompute or adjust weights.
- Save the character into your rigged-character library.
- Send the rigged character to Refine or Animate.
For the most reliable workflow, save from Rig after building the rig. Once the character is saved as a rigged character, Refine can persist weight and texture edits for that character.
If you open a generated or uploaded model that already has a saved rig, Create or Uploads may show actions to edit the rig, refine it, animate it, or discard the saved rig and return to the original source model.
Refine
The Refine tab is for already-rigged characters. It is locked until the current character has skeleton and weight data.
Use Refine to improve a rig after the first pass. The workspace can save updated weights and, when texture painting or texture output is available in the refinement session, save the updated texture as part of the character.
Important workflow note: if you rig a new model and go directly to Refine without saving the rigged character first, your current edits can remain available during the live session, but cloud-persistent Refine saving requires the character to exist in your rigged-character library. If Refine tells you the character is not in your library yet, return to Rig and use Save Rigged Character first.
Animate
The Animate tab is for posing and animating rigged characters. It is locked until the current character has a rig.
You can open Animate from:
- A newly rigged character in Rig.
- A refined character in Refine.
- A rigged generation in Create.
- A rigged upload in Uploads.
- A saved character from your rigged-character library.
If you try to use Animate while the current model is not rigged, 3D Studio sends you back to Rig so you can complete the setup.
Animate can also hand the current character back to Refine, which is useful when you test a pose or motion and notice weighting problems that need cleanup.
Import
The Import tab is for bringing in supported pre-rigged character files and opening them directly in Animate with their original skeleton and skinning preserved. Imported characters are treated differently from generated/unrigged models: their existing rig and textures are kept intact, and the standard Rig and Refine pipeline is not used for them.
Use Import when you already have a rigged character from another character tool and want to pose or animate it in 3D Studio rather than generate and rig a new one from scratch.
Rigged characters and the current-character workflow
3D Studio is organized around one current character. When you click a generated model, uploaded model, imported character, or saved rigged character, the studio promotes it to the current character. The current-character bar below the tabs shows:
- Character name.
- Thumbnail, when available.
- Whether it is unrigged or rigged.
- Basic mesh or rig metadata, such as bone and vertex counts when available.
- Hints such as loading activity or unsaved changes.
The current character follows you across compatible tabs. For example, if you select a generated model in Create, then click Rig, that model is loaded for rigging. If the selected model already has a saved rig, Refine and Animate can open the rigged version.
Although the older dedicated Rigged Characters library workflow may still appear in some contexts, the main Create gallery now includes Rigged and Unrigged filters so you can manage generated characters without leaving Create. Saved rigged characters can be opened for animation, edited in Rig, refined, or deleted where the library view is available.
Tab availability
3D Studio enables tabs based on what is currently open.
If no model is open:
- Rig is disabled.
- Refine is disabled.
- Animate is disabled.
If an unrigged model is open:
- Rig is available.
- Refine is disabled until the model has rig and weight data.
- Animate is disabled until the model has a rig.
If a rigged character is open:
- Rig, Refine, and Animate are available.
If an imported pre-rigged character is open, the studio preserves its original rig and focuses the workflow on Animate rather than running it through the generated-character rigging pipeline.
Some experimental or restricted tabs may not appear for all users.
Unsaved changes
3D Studio tracks unsaved edits in Rig and Refine. If you try to switch from the current character to a different character while either tab has pending edits, the studio warns that switching will discard those changes.
Choose Cancel if you need to go back and save first. Choose Continue only if you are comfortable losing the unsaved edits for the previous character.
Switching tabs for the same character does not automatically discard edits. The warning is specifically about changing to a different character while unsaved Rig or Refine work is pending.
Typical workflows
Generate an image-based 3D character and rig it
- Open Create.
- Choose Image input mode.
- Upload a clean front-facing character image. An A-pose or T-pose style image is best.
- Select a generation model.
- Enable rig-friendly pose settings if the selected model offers them.
- Generate the model.
- Open the completed model in the viewer.
- Inspect the mesh and texture.
- Send it to Rig.
- Build or adjust the rig.
- Save the rigged character.
- Open Animate to test it, or Refine to improve weights.
Generate multiple variations from a batch of images
- Open Create and choose Image.
- Drop several images into the upload area.
- Optionally label each staged image.
- Select the generation model or models you want to test.
- Start the batch.
- Review the resulting gallery cards.
- Use search, rename, and reuse-source tools to organize the best results.
Use several reference views
- Open Create and choose Multi-Img.
- Upload 2–4 images of the same subject.
- If view slots are shown, fill Front first, then add side/back views.
- Select a compatible generation model.
- Generate and inspect the result.
- Rig the model if you plan to animate it.
Continue work on a rigged generation
- Open Create.
- Switch the gallery filter to Rigged.
- Choose the character.
- Use the available actions to edit the rig, refine it, or animate it.
Start over from the original mesh
- Find the generated or uploaded source that already has a saved rig.
- Use the discard/revert rig option when shown.
- Reopen the original unrigged source.
- Send it to Rig again.
- Save the new rigged character.
Tips and troubleshooting
I want a 3D model from an image. Which tool should I use?
Use 3D Studio. Open Create, choose Image, upload your image, select a generation model, and generate. Use Voxel Studio only if you specifically want a blocky voxel/cube model.
My character is hard to rig
Use a clearer source image. Best results usually come from a front-facing character with visible limbs, a neutral standing pose, and separation between arms and torso. Avoid crossed arms, seated poses, heavy props blocking the body, or extreme camera angles.
Refine or Animate is disabled
The current model is not rigged yet. Open it in Rig, build or load a rig, and save or send the rigged character forward.
Rig is disabled
No model is currently open. Select a completed generation in Create, choose a model in Uploads, or open a saved rigged character.
The viewer is slow to load
Large 3D files can take time to download and prepare. Watch the loading meter. If the model is very dense, use the optimization tools after it loads and compare the optimized version before rigging or exporting.
My optimized model looks too different
Try a less aggressive optimization target. Use the Both comparison view to check silhouette, texture clarity, and important details before using the optimized mesh downstream.
Refine says edits cannot be saved
The character may not be saved as a rigged character yet. Return to Rig, click Save Rigged Character, then go back to Refine and save again.
I switched characters and my edits disappeared
If you confirmed the unsaved-changes warning, pending Rig or Refine edits for the previous character were discarded. Save important changes before switching to another character.
A generated model already shows as rigged
That means the source has a saved rigged version. Use the available actions to edit the rig, refine it, animate it, or discard the rig and return to the original mesh.
FAQ
Does 3D Studio make real 3D mesh files?
Yes. 3D Studio produces smooth polygon mesh models with textures when the selected generation settings create them. These are different from voxel cube models.
Can I animate an unrigged generated model?
No. Animation requires skeleton and skin-weight data. Send the model to Rig first.
Can I download the generated model before rigging?
Yes, when downloadable model files are available for the selected generation. Open the completed generation and use the download options in the inspector panel.
Can I use uploaded models?
Yes. Open Uploads, select a model, and send it through the same Rig, Refine, and Animate workflow.
Can I import an already-rigged character?
Yes, supported pre-rigged imports can be opened through Import and sent to Animate while preserving their original rig and textures.
Why does clicking Create again take me back to the gallery?
When you are already on the Create tab, clicking the active Create tab resets the Create view back to the main gallery grid. This is a quick way to leave a model detail view.