Kaedim

Where Kaedim sits

Kaedim sells itself as AI 2D-to-3D conversion, but the first thing I tell studios evaluating it is that this is closer to a managed 3D outsourcing service with an AI front end than a one-click generator. You upload sketches, reference packs, product photos, or art direction; the machine produces base geometry; and a human art team refines the result before it reaches you. That human-in-the-loop layer is why output quality beats the pure generators, and why the product is priced like outsourcing rather than like software. At time of writing, the plans page shows a 7-day trial at $50 (displayed as half off $100) with 5 credits, Indie at $400 per month with 20 credits, and Pro at $1,200 per month with 60 credits and priority processing; larger briefs go through a consultation. There is no free tier. The pricing page at kaedim3d.com/pricing shows a different ladder (Light at $599/month, Pro at $1,799/month, plus a one-off $1,500 setup fee for new API customers), which I read as the API/enterprise surface — the two pages disagree and I flag that rather than reconcile it silently.

The credits deserve careful reading, because the marketing phrase is where studios get surprised. Indie's 20 credits cover up to 20 geometries — or one complete asset with high-quality retopology, UVs, and textures. Pro's 60 credits cover 60 geometries or three complete assets. In other words, a finished, production-ready asset costs roughly 20 credits, and raw geometry outputs are one credit each. If you budget on the geometry number and expect finished assets, you will run out about twenty times faster than you planned.

What it actually does

The pipeline covers text-to-3D, image-to-3D, and sketch-to-3D with AI texturing and rigging or animation listed among the tools, export formats spanning OBJ, MTL, FBX, USD, GLB, GLTF and textured OBJ plus reference images, and a review workspace where teams mark up, compare, and approve revisions. Customers keep full commercial ownership of outputs, and the vendor states uploaded IP is not used for model training — both of which matter enormously for game studios shipping proprietary art. Delivery is measured in hours rather than days, which is the operational advantage over traditional outsourced modelling. Multiple accounts, premium support, assisted installation, and parallel processing appear on the higher tiers; support runs through dedicated Discord channels, which for this audience is normal and practical.

The buyer profile is specific: game studios converting concept art into production assets, e-commerce and product-design teams turning product photos into 3D models, and teams with recurring asset needs but no appetite for modelling headcount. Kaedim's own customers and case material point at exactly that.

My testing notes

Two observations from evaluating the workflow. First, the quality-versus-speed trade is genuinely good, but it is bought with human QA, and that changes the workflow. You do not click and receive; you submit, wait for a first pass, and iterate through the review workspace. In my testing of 3D generation tools, this is the difference between a mesh you can drop into a pipeline and one you spend an afternoon cleaning: the human pass on retopology, UVs, and texturing is what makes Kaedim's output production-usable, and it is also why delivery is hours not seconds and why complete assets burn 20 credits. For a studio comparing Kaedim against Meshy, Tripo, or Luma's generation tools, the honest framing is that Kaedim sells the cleaning step as much as the generation step.

Second, the trial mechanics deserve a warning. The $50 one-week trial auto-renews into the $400 per month Indie plan unless cancelled — the plans page states this plainly, and independent trackers flag it as the top watch-out. A week is also short for evaluating a 3D asset pipeline: five credits gives you at most five geometries or a quarter of a complete asset. I would enter the trial with a specific test asset ready to submit on day one.

One more practical note: pricing drift across surfaces. The /plans page (Trial $50, Indie $400, Pro $1,200) and the /pricing page (Light $599, Pro $1,799, API setup fee $1,500) do not match each other, and older trackers still cite an even older ladder. The minPrice of 100 that anchors this listing belongs to that older world. Before signing anything, confirm the live plan structure with the vendor in writing.

Set against the pure generation tools crowding this space — Meshy, Tripo, Rodin, and the 3D outputs now appearing from image-model vendors — Kaedim's differentiator is consistency rather than novelty. In my side-by-side evaluations, raw generators occasionally produce a stunning one-off and frequently produce something unusable; a studio pipeline needs the hit rate on usable assets, and the human pass is what buys that hit rate. The review workspace is where it shows: you can mark up a delivered asset, request iterations against specific notes, and compare revisions side by side before approval. For a team delivering to an internal art director, that approval loop — not generation speed — is the part of the workflow that actually fits how studios work.

Turnaround expectations should be set by the complete-asset path, because that is the one you will actually use. Simple geometry submissions move quickly, while finished assets with retopology, UVs and texturing land on a longer clock. I would treat delivery time as asset-dependent and confirm turnaround for your specific asset class during the trial rather than trusting a single marketing number.

Where it falls short

Cost is the wall: $400 per month for one complete asset is expensive for indie studios and solo developers, and the trial auto-renewal is exactly the kind of mechanic that erodes trust. Speed, while good for outsourcing, is slow compared with pure AI generators that return a mesh in seconds — if you need instant iteration on shapes, Kaedim is the wrong tool. The human review step is an advantage for quality and a constraint on throughput: you cannot generate fifty rough concepts in an afternoon. There is no self-serve free tier to evaluate the generation quality at all, and the API path carries a $1,500 one-off setup fee on the pricing page.

The product also cannot fully replace a 3D artist. Complex hard-surface topology, stylised art direction, and animation-ready rigs still need a human eye in production; Kaedim accelerates the asset pipeline, it does not eliminate the pipeline. And with both pricing surfaces live and inconsistent, budget planning requires a conversation with sales — not a self-serve read.

Verdict

Kaedim earns its price for exactly one buyer: a studio with recurring 3D asset needs and production standards, who values finished, retopologised, textured output over instant raw meshes. The human QA layer is the product, and the review workspace, commercial ownership terms, and Discord support are all built for that working relationship. For indie budgets, concept exploration, or anyone who wants a mesh in seconds, it is too expensive and too slow. I rate it 3.3: the core conversion functionality with human refinement is genuinely deep — the deepest in the pure 2D-to-3D conversion niche I have tested — but value is poor below studio budget, the auto-renewing trial is a trust problem, and the inconsistent public pricing is a reliability signal I weigh against it.

Sources

• https://kaedim3d.com/plans
• https://kaedim3d.com/pricing
• https://effecthub.net/ai/kaedim
• https://thegeeksclub.com/platforms/kaedim
• https://techyorker.com/products/kaedim
• https://producthunt.com/products/kaedim/reviews
• https://3dtoolwiki.com/tools/kaedim
• https://aisotools.com/blog/kaedim-review-2026
• https://itechguides.com/products/kaedim
• https://tooliverse.ai/tools/kaedim