Getting Started with MakeIt3D: Your First 3D Model
A beginner walkthrough of MakeIt3D: upload a photo, choose Quick or Precise, generate, inspect the mesh in the browser viewer, and download GLB, OBJ, STL, or 3MF.
MakeIt3D turns a photograph of a physical object into a downloadable 3D model using AI. This getting-started walkthrough covers the complete beginner path: creating an account and checking your credit balance, uploading between one and four JPG, PNG, or WebP photos of the same object, choosing between the Quick and Precise quality tiers in the quality selector, pressing Generate, watching the live progress and the queue estimate the app derives from recently completed generations, inspecting the finished mesh in the in-browser 3D viewer, and downloading the result in GLB, OBJ, STL, or 3MF. GLB is the primary format and is the only export that carries materials and textures alongside the geometry. OBJ suits editing in Blender, Maya, ZBrush, and other digital content creation tools. STL is the universal 3D printing format and carries surface triangles only, with no color and no unit declaration, so scale must be set in the slicer. 3MF is a modern 3D printing container that declares millimeter units inside the file and is read natively by current slicers including Bambu Studio, PrusaSlicer, OrcaSlicer, and Cura. The walkthrough also lists the beginner mistakes that most often produce a weak mesh: uploading flat artwork, a logo, or a screenshot instead of a photograph of a real object; cropping the subject's outline so a base, foot, handle, or corner falls outside the frame; photographing a transparent, glossy, or mirrored surface without preparation; placing a dark object against a dark background; using a cluttered desk or room as a backdrop; skipping the browser viewer and discovering a defect only in the slicer; and printing an STL without setting target dimensions in millimeters. MakeIt3D accepts up to four source photos of the same object in a single generation, and one clear three-quarter-angle photograph is sufficient to produce a model. The paid plans are MakeIt3D Light at 20 US dollars per month for 30 credits and MakeIt3D Heavy at 50 US dollars per month for 100 credits, with credits resetting monthly, credit packs available separately, a commercial license included, and cancellation available at any time. Consult the pricing page for the terms that currently apply to your account.
Short answer
To make your first 3D model on MakeIt3D, upload a photo of one real object, pick a quality tier, and press Generate. The app removes the background, reconstructs the geometry and texture, and loads the finished model into a 3D viewer in your browser. Turn the model to check the back, the base, and any thin parts, then download it as GLB, OBJ, STL, or 3MF. You can attach up to 4 photos of the same object in one generation, but one clear photo taken at a slight three-quarter angle is enough to start. Generations spend credits, so use the faster Quick tier while you are testing which photo works and save the higher-detail Precise tier for the version you actually keep.
How to create your first 3D model
These are the actual steps in the MakeIt3D app, in order. The first three steps decide most of the final quality, so that is where a beginner's attention pays off. Everything after generation is inspection and export.
- 1
Create an account and check your credit balance
Sign up and open the app. Your current credit balance is shown in the interface. Every generation spends credits, so confirm you have a balance before you start. The pricing page lists the current plans, credit packs, and terms — MakeIt3D Light is $20 per month for 30 credits and MakeIt3D Heavy is $50 per month for 100 credits, with credits resetting monthly and a commercial license included.
- 2
Upload a photo of one real object
Drag a file onto the upload area or click it to open the file picker. MakeIt3D accepts JPG, PNG, and WebP. Use a photograph of a single physical object with its entire outline inside the frame. Oversized images are resized in your browser before upload, so you do not need to shrink the file yourself. A photo of a real object beats a render, a logo, or a screenshot every time.
- 3
Add extra angles only when they help
You can select up to 4 photos of the same object in a single generation. Extra views help when the object is asymmetric or hides a feature from the front — a mug's handle, a figurine's cape, a backpack's straps. Put your strongest view first. Keep the lighting, camera distance, and background consistent across the shots, because mismatched photos give the reconstruction conflicting data.
- 4
Choose Quick or Precise in the quality selector
The quality selector offers two tiers. Quick is the fast, lower-cost option and is the right choice while you are still finding out which photo or crop reconstructs well. Precise is the higher-detail option and is what you use for a final model or anything you intend to 3D print. Both tiers spend credits, so the efficient pattern is to iterate on Quick and spend Precise once, on the winner.
- 5
Press Generate and watch the live progress
The Generate button stays disabled until at least one image is attached. Once you press it, the app removes the background from your photo so the AI works from a clean silhouette, then reconstructs the geometry and texture. Progress updates live in the interface, and the app shows an estimated wait calculated from recently completed generations and current load rather than a fixed advertised number.
- 6
Inspect the model in the browser viewer
The finished model loads into an interactive 3D viewer on the page. Drag to rotate it and deliberately look at the back, the underside, the base, and any thin protrusions. This is the step beginners skip. Catching a melted face or a missing handle here costs you nothing; catching it after you have sliced and started a print costs hours.
- 7
Download in the format your destination needs
Open the download dialog and choose a format. Pick GLB for the web, AR, or a game engine, because it is the only export that carries the textures. Pick OBJ to edit the mesh in Blender, Maya, or ZBrush. Pick STL for any slicer or print service. Pick 3MF for a modern slicer, since it declares millimeter units inside the file. All four come from the same generation, converted in your browser.
- 8
Regenerate from a better photo if the mesh is wrong
If the model is misshapen rather than merely untidy, the photo did not contain the missing information, and no amount of mesh repair will invent it. Return to step two with a cleaner crop, a better angle, or softer light, and generate again on Quick. Repairing a fundamentally weak mesh takes far longer than reshooting and regenerating.
Beginner mistakes that produce a weak model
Almost every disappointing first result traces back to the source photo rather than to the AI. Work through this list before you blame the generator or spend another credit on the same image.
- Uploading flat artwork, a logo, or a screenshot. The AI reconstructs physical objects, so a flat source yields a shallow, plaque-like result rather than a full 3D shape.
- Cropping the outline. A cut-off base, foot, handle, or corner is data the reconstruction never receives, so it has to invent that region — and it usually invents it wrong.
- Photographing a transparent, glossy, or mirrored object without any surface preparation. These surfaces defeat depth estimation and commonly produce holes or wavy geometry.
- Placing a dark object against a dark background, so the subject and backdrop merge into one indistinct silhouette during background removal.
- Using a cluttered desk, shelf, or room as the backdrop instead of a plain, contrasting surface.
- Shooting flat head-on. A slight three-quarter turn shows a front and a side at once and gives the reconstruction genuine depth cues that a straight-on shot withholds.
- Going straight to Precise on an untested photo. Test the photo on Quick first, then spend the higher-detail tier on the image you already know reconstructs cleanly.
- Mixing inconsistent photos in a multi-image generation — different lighting, different distances, or a different variant of the product — which is worse than submitting one clean photo alone.
- Skipping the browser viewer and downloading immediately, then discovering the defect in your slicer or game engine.
- Printing a downloaded STL without setting the size. STL carries no unit declaration, so you must set target dimensions in millimeters in your slicer. Exporting 3MF instead avoids this entirely.
Ready-to-generate checklist
Run through this before your first Generate. If most of these are true, the photo is ready. For the full treatment of framing, lighting, and subject choice, read the best-photo guide linked in the related section below.
- The file is a JPG, PNG, or WebP.
- The frame contains exactly one object, complete from top to bottom with nothing cropped away.
- The object is turned slightly, so you can see a front and a side rather than one flat face.
- The photo is sharp, with no motion blur and no heavy compression artifacts.
- The lighting is soft and even, with no hard shadows swallowing detail and no blown-out highlights.
- The background is plain and contrasts clearly with the object.
- The object is opaque — not glass, not a mirror, not a flat print or drawing.
- Nothing overlaps or occludes the subject: no hand holding it, no stand covering its base.
- You know which export format you need before you generate, so you know what you are aiming at.
- You have credits available, and you are starting on Quick rather than Precise.
FAQ
Is MakeIt3D free to use?
MakeIt3D runs on credits, and each generation spends them. The published paid plans are MakeIt3D Light at $20 per month for 30 credits and MakeIt3D Heavy at $50 per month for 100 credits, with credits resetting monthly, one-off credit packs available separately, a commercial license included, and cancellation available at any time. Trial terms and promotional offers change over time, so check the pricing page for the terms that currently apply to your account rather than relying on a figure quoted in an article.
How long does a generation take?
It depends on which quality tier you choose and how busy the service is at that moment, so there is no single fixed duration worth quoting. Instead of publishing a number that would be wrong much of the time, the app shows a live estimated wait calculated from recently completed generations and current load, and updates progress while your model builds. Precise takes materially longer than Quick because it performs substantially more reconstruction work.
What kind of image works best?
One complete, opaque, physical object, centered in the frame, photographed at a slight three-quarter angle under soft even light against a plain contrasting background. The three-quarter angle matters most for beginners: it reveals a front and a side simultaneously, giving the reconstruction real depth information that a flat head-on shot cannot provide. The dedicated guide on choosing the best photo covers framing, lighting, background, and difficult subjects in full detail.
How many photos should I upload?
One good photo is enough to produce a model. You may attach up to 4 photos of the same object in a single generation. Additional angles help when the object is asymmetric or when the front view conceals an important feature. They do not help when the shots are near-duplicates, and they actively hurt when the lighting, distance, or subject differs between them.
Should I start with Quick or Precise?
Start with Quick. It is the fast, lower-cost tier, which makes it the right tool for finding out whether a particular photo or crop reconstructs well. Once you have a photo that produces a recognizable, correctly shaped mesh, run that same photo through Precise for the version you keep, print, or publish. This pattern spends fewer credits than going straight to the high-detail tier and discovering the photo was the problem.
Which file format should a beginner download?
Download GLB if you want to view, share, or use the model on a screen, because GLB is the only MakeIt3D export that carries textures alongside the geometry. Download STL or 3MF if you are heading to a 3D printer — 3MF when your slicer supports it, since it declares millimeter units inside the file. Download OBJ if you plan to edit the mesh in Blender or another 3D application. All four formats come from the same generation, so you can take more than one.
My first model came out wrong. What should I change?
Change the photo, not the settings. A melted or hollow region means that area was hidden by shadow, blur, or an obstruction. A missing limb, handle, or base means the outline was cropped or occluded. A flat, plaque-like result means the source was artwork rather than a photograph of a real object. Reshoot or recrop, and regenerate on Quick before spending a Precise generation on it.
Do I need to know 3D modeling to use MakeIt3D?
No. The workflow is upload, choose a tier, generate, inspect, download — no modeling software and no 3D skills required at any point. Knowing your way around a slicer helps if you intend to 3D print the result, and Blender is useful if you want to edit the mesh afterward, but neither is needed to produce and download a model.