Kaedim Review 2026: Pricing, Features, Pros & Cons
Kaedim turns a piece of concept art into a game-ready 3D mesh — using a model for the first pass and human artists for the part models are still bad at. Here's an honest look at what that hybrid actually buys you, what the credit pricing works out to per asset, and where it loses to a pure generator in 2026.
Quick Verdict
Best for: studios with a steady list of props, environment pieces and set dressing who currently outsource that work and want it back faster and cheaper. Skip it if: you want free fast iteration on ideas, or you are buying it believing no humans are involved — both readings will leave you unhappy.
What Is Kaedim?
Kaedim is a 2D-to-3D service for game and interactive content teams. You upload concept art, a reference photo or a flat image, specify the poly budget and the target engine conventions, and you get back a 3D model built to that spec. The distinction it draws with the rest of the category is production-readiness: the promise is not "a mesh" but "a mesh your technical artist will accept", with topology, UVs and poly counts that do not require a full retopology pass before the asset can be used.
The mechanism behind that promise is the thing that makes Kaedim controversial and also the thing that makes it work. A model generates the first pass; human 3D artists then refine it. In 2026 the company is open about this, but the early framing was aggressive enough that a chunk of the 3D community still describes the product as outsourcing with a machine-learning veneer. That framing is unfair to the model and unfair to the artists, but the practical implication is real: you should evaluate Kaedim against your outsourcing costs and turnaround, not against a text-to-3D generator's per-minute speed.
Turnaround is measured in hours rather than minutes, revisions are a request rather than a re-roll, and pricing is a monthly plan with credits consumed per generation. That shape tells you the buyer it is designed for: not a hobbyist exploring an idea, but a team with a spreadsheet of assets that has to exist by a milestone date.
Built a 3D, game-asset or generative modelling tool? People land on this review while they are still choosing one.
Add it to the 3D category — a free listing publishes after review, and it is the same page ChatGPT, Perplexity and Google read when a studio asks which image-to-3D tool to use. Want it live in minutes with a Verified badge instead? That option is on the form, one-time, no subscription.
Kaedim Pros & Cons
✓ Pros
- •Output arrives as a usable mesh rather than a photogrammetry-style blob — quad-dominant topology and sane UVs are the point of the product, and that is the part most pure-generative tools still get wrong
- •The human-in-the-loop step means a failed generation is somebody's problem rather than yours: you request a revision instead of re-rolling a prompt and hoping
- •It is aimed squarely at production pipelines — the export formats, the poly-count targets and the asset conventions assume the mesh is going into a game engine, not a demo reel
- •Turnaround measured in hours beats an outsourced modelling brief measured in days, which is the honest comparison for the studios buying it
- •Consistency across a set of assets is better than prompt-driven generators manage, because a person is enforcing it rather than a seed
- •Removes the most tedious tier of work — background props, set dressing, kitbash pieces — which is exactly where a small team's modelling hours quietly disappear
✗ Cons
- •Credit pricing makes per-asset cost hard to predict up front, and a complex model can burn several credits before it is right — teams routinely underestimate the real per-asset number
- •It is not instant, and it is not self-serve in the way the marketing implies: the artist step is what makes the output good and it is also what puts a queue between you and the mesh
- •The company's early positioning drew sustained criticism from 3D artists for how much of the work was human rather than model, and if you are buying it as pure AI you are buying the wrong description of it
- •Hard-surface and mechanical shapes come back better than organics; characters, cloth and anything with fine anatomical detail still need real cleanup
- •No free tier worth evaluating on — you are committing to a paid plan before you know whether your art style survives the pipeline
- •Vendor lock at the pipeline level is real: the workflow assumes their web app as the entry point rather than a plugin inside the DCC tool your artists already live in
Kaedim Pricing 2026
Kaedim sells monthly plans carrying a credit allowance rather than a per-model price, and complexity determines how many credits an asset consumes. That makes the headline number a poor guide to what you will actually spend — the only reliable estimate comes from running one real month of your asset list through it and dividing.
Starter
- •Monthly credit allowance
- •Standard turnaround queue
- •Common game-engine exports
- •Web app access
Solo devs and small studios testing whether the output survives their style
Pro / Studio
- •Larger credit pool
- •Priority queue
- •Team seats
- •Revision handling
Studios pushing steady prop and environment volume
Enterprise
- •Volume credit pricing
- •Pipeline and API integration
- •Dedicated support
- •Custom asset specs
Publishers with an ongoing outsourcing budget to redirect
Tier shape as publicly described by Kaedim as of September 2026; exact allowances and figures move and higher tiers are quote-based, so confirm current numbers before budgeting. Credits are consumed per generation, and revisions on a complex asset can cost more than the first pass did.
Kaedim vs Meshy vs CSM AI
| Feature | Kaedim | Meshy | CSM AI |
|---|---|---|---|
| Input | 2D image or concept art | Text or image | Image, video or text |
| Human refinement step | ✅ Core to the product | ❌ Fully generative | ❌ Fully generative |
| Turnaround | ⚠️ Hours, queued | ✅ Minutes | ✅ Minutes |
| Game-ready topology | ✅ The main selling point | ⚠️ Improving, still needs retopo | ⚠️ Varies by asset |
| Free tier | ❌ None meaningful | ✅ Free credits | ✅ Free credits |
| Entry price | ⚠️ Studio-level monthly | ✅ Low monthly | ✅ Low monthly |
| Revisions | ✅ Request a fix | ⚠️ Re-generate | ⚠️ Re-generate |
| Best fit | Props and environments at volume | Fast iteration and prototyping | Scanning real objects into assets |
When Kaedim Is the Right Call
The clearest case is a studio that already outsources prop and environment modelling. That team has a known cost per asset, a known turnaround in days, and a known coordination overhead in briefs and revision emails. Kaedim competes on all three of those numbers rather than on novelty, and when it wins it usually wins on turnaround first — a milestone that needed forty pieces of set dressing does not slip because a vendor was booked.
The second case is a small team whose art director is also its only modeller. The bottleneck there is not skill, it is hours, and almost all of the hours go to assets nobody will look at closely. Handing that tier to a service that returns engine-ready geometry frees the one person who can actually set the visual direction to do it.
Where it is the wrong call is exploration. If you are still deciding what a thing looks like, you want to generate twenty variations for nothing and throw away nineteen — that is a generative tool's job, and paying credits plus a queue for each attempt is the wrong economics entirely. It is also the wrong call for hero characters, where the cleanup and art-direction loop tends to consume whatever the pipeline saved.
Frequently Asked Questions
Is Kaedim actually AI, or is it people?
It is both, and the company has been clearer about that over time than it was at the start. A model produces the first pass from your 2D input, and human 3D artists then clean up topology, fix geometry the model got wrong and bring the asset to the spec you asked for. That hybrid is the reason the output is usable in an engine without a retopology pass, and it is also the reason it costs more and takes longer than a purely generative tool. If your buying criterion is 'AI did all of it', Kaedim will disappoint you. If your buying criterion is 'a game-ready mesh landed in my pipeline this afternoon for less than an outsourcing brief', the hybrid is a feature.
How much does Kaedim cost per model?
There is no clean per-model number, which is the most common complaint from teams evaluating it. You buy a monthly plan carrying a credit allowance, and each generation consumes credits based on complexity — a simple prop is cheap, a detailed character is not, and a model that needs several passes multiplies. Entry plans start in the low hundreds per month. The honest way to evaluate it is to take one month of your real asset list, run it, and divide; teams that do this usually land somewhere well above their initial estimate but still below what the same list would cost from an outsourcing studio.
Kaedim vs Meshy — which should I use?
They are answering different questions. Meshy is fast, cheap, self-serve and fully generative: you get a mesh in minutes, you iterate freely, and you accept that topology often needs work before it ships. Kaedim is slower, more expensive and partly human: you get an asset that is closer to production spec and you can ask for a revision when it is not. Prototyping, blockouts and 'what would this look like in 3D' work favour Meshy. Filling a shipping game's prop and environment list favours Kaedim. Plenty of studios use the fast tool for exploration and the hybrid one for the assets that actually go in the build.
What kind of assets does Kaedim handle best?
Hard-surface and inorganic shapes — props, weapons, furniture, crates, architectural pieces, set dressing — come back most reliably, because the geometry is describable and the cleanup is bounded. Organic subjects are the weaker end: characters, creatures, cloth, hair and anything where an artist's eye is judging anatomy rather than tolerance. That maps neatly onto where the value is anyway, since the tedious volume in most productions is props and environments, not hero characters, which studios rarely want to hand to any external pipeline.
Does Kaedim have a free trial?
Not in any form that lets you evaluate it properly. Unlike the generative competitors, which hand out free credits so you can throw ten images at them in an evening, Kaedim's model is a paid plan from the start. That raises the cost of finding out whether your specific art style survives the pipeline, and it is the main reason to run a deliberate one-month trial with a real asset list rather than a casual look. Ask about a pilot before committing to an annual arrangement.
Can Kaedim replace a 3D artist?
No, and the product's own architecture is the evidence — there are human 3D artists inside the service doing the refinement. What it replaces is a specific slice of a 3D artist's week: the low-creativity, high-volume modelling that has to exist for the world to feel populated. Teams that get value from it redirect their in-house artists to hero assets, style direction and the things a pipeline cannot judge. Teams that expect to cut headcount and keep the same output usually find the revision loop and the art-direction gap eat the savings.
Compare AI 3D Tools
See how Kaedim stacks up against Meshy, Luma AI and every other 3D generation tool in the directory.
Affiliate disclosure: Some links on this page are affiliate links. If you sign up through them, AISO Tools may earn a commission at no extra cost to you. This never affects our rankings or reviews.
📬 Get the best new AI tools delivered weekly
One concise email with fresh launches, trending picks, and featured standouts.
Join thousands of professionals who discover the best AI tools every week. No spam — unsubscribe anytime.