A concept sketch, a product photo, a clay figure on your desk upload it, and about a minute later there's a textured, exportable 3D mesh. Here's the complete guide to Meshy's Image to 3D: how it works, what it costs per conversion, the input-photo tricks that separate great results from mush, and its honest limits.
⚡ Up to 8 images per conversion on Meshy 6 · ~60s generation · free retries for Image to 3D on paid plans
1–8 images of your subject. One works; several angles of the same object work far better (more on that below).
Pick the model series (Meshy 6 for max detail), texture on/off, target polycount and for characters, pose control: A-pose, T-pose or custom.
The AI reconstructs geometry in ~60 seconds. Multiple variants let you pick the best interpretation.
Texture it (full PBR maps), run Smart Remesh for clean topology at your polycount, or re-roll Image to 3D retries are free on paid plans.
Download in 6 formats, pull it through an engine/slicer plugin or keep going: auto-rig and animate humanoid characters right there.
The classic pipeline: character sheets and prop concepts become meshes that actually match the art direction then flow into rigging. A/T-pose art converts cleanest.
Your own products become AR previews, e-commerce spinners and marketing renders. Shoot multiple angles for accurate geometry all round.
Clay maquettes, sculpts, toys, found objects photograph from several sides, convert, then STL out to your printer (see our printing guide).
Clean line art with clear silhouettes converts surprisingly well; loose, ambiguous or heavily-shaded sketches force the AI to guess. One clear subject per image.
Midjourney/DALL·E art converts like concept art and inconsistent lighting or impossible geometry in the source becomes weirdness in the mesh. Pick renders with plausible, solid forms.
Strong overall with the known caveat that faces and hands are AI generation's weak spot. Multi-view input and a re-roll budget (free on paid plans for this feature) get you there.
From a single front photo, the model must invent the back and sides of your object. Sometimes brilliantly, sometimes… creatively. Meshy 6's multi-view input (up to 8 images) replaces that guesswork with evidence:
Fast and effortless right for simple, roughly symmetrical objects, or when only the front matters (relief-style decor, background assets). Expect the unseen side to be plausible rather than accurate.
Front, back, sides each angle constrains the reconstruction. The difference shows most on asymmetric objects, characters, and anything you'll inspect from all sides (prints, AR products, game assets).
Quick capture recipe for real objects: place the object on a plain surface, walk around it shooting every ~45–90°, keep lighting even and the framing consistent. Eight decent phone photos beat one perfect one.
What good input looks like.
The classic input mistakes.
Same honesty standard as our full review know these before you rely on the feature:
| Path | Credits available | ≈ Textured Meshy 6 conversions (30 cr) | Notes |
|---|---|---|---|
| Free plan | 100 / month, $0 | ~3 / month | 10 downloads/mo (Meshy 5 models); assets public, CC BY 4.0 |
| Pro $20/mo ($10 first month) | 1,000 / month | ~33 / month (≈ $0.60 each) | Meshy 6 downloads, private ownership, free Image-to-3D retries, 10 concurrent tasks |
| Draft workflow trick | Any plan | Up to 6× more drafts | Use non-flagship series (5 cr mesh-only) for exploration; spend Meshy 6 credits on keepers |
Full credit mechanics on the pricing guide · automating conversions? The same per-call costs apply via API see API pricing.
Upload 1–8 images; Meshy's AI reconstructs full 3D geometry from them in about a minute, optionally generating PBR textures (base colour, metallic, roughness, normal). Multiple angles of the same subject constrain the reconstruction for accurate sides and back; pose control (A/T/custom) prepares characters for rigging. The mesh then exports in six formats or continues into Meshy's remesh → rig → animate pipeline.
Yes 100 credits monthly on the free plan, no card, which covers several conversions (a textured Meshy 6 conversion runs 30 credits; other series as little as 5 mesh-only). Downloads on free are limited to 10/month of Meshy-5 models; paid plans unlock Meshy 6 downloads, private ownership and free Image-to-3D retries. Details on our free plan guide.
No, and the difference matters: photogrammetry measures dozens/hundreds of photos, dimensional fidelity, real surface detail; AI image-to-3D interprets a few images, ~60 seconds, and a plausible model that looks right. Scanning wins for accuracy-critical work; Meshy wins for speed, for working from art (which can't be scanned), and for objects you only have one picture of. Many workflows use both.
Yes stylised art is a core use case, and clean character sheets with clear silhouettes convert best. Two cautions: heavy shading gets read as geometry, and converting other people's copyrighted characters means you can't commercially own the result regardless of how well it converts the licence rules hinge on your inputs.
Because the AI never saw it from one photo, unseen surfaces are educated guesses. The fix is multi-view input: shoot your subject every 45–90° (Meshy 6 takes up to 8 images) and the guessing largely disappears. For irreducibly single-image sources like a painting, accept the invented back or sculpt-correct it in Blender.
Sharp and well-lit matters more than enormous standard phone-camera resolution is plenty. Meshy accepts uploads up to 50MB. Spend effort on lighting, background and multiple angles rather than megapixels; a crisp 12MP set beats one soft 100MP frame every time.
Use Image to 3D when the design already exists concept art, products, real objects and fidelity to that source is the goal. Use Text to 3D when you're exploring from imagination and iteration speed beats reference-matching. They combine well: generate concepts by text, pick a favourite, then image-convert your refined 2D art for the final asset.
Upload it free 100 monthly credits, no card, about a minute per conversion. Bring a few angles and watch the difference.
Convert Your First Image Free →