Reference and specification intake
Submit up to ten views of one object plus dimensions, polygon targets, topology, mesh groups, UV, texture, rig and LOD requirements. Clear orthographic references reduce ambiguity.
O05O08O10Kaedim
A managed AI-plus-artist route from references and specifications to reviewable 3D assets—not a cheap instant generator.
Evidence statusPublic-evidence reviewMedium confidenceLast checked Sep 12, 2026By MakeGameWithAI
PRODUCT OVERVIEW
What Kaedim is—and how it works.
Kaedim is best understood as a software-managed 3D production service. A team submits sketches, reference packs, product images, briefs, art bibles, technical specifications and art direction; Kaedim combines proprietary AI with an internal art team or Art Lead, then gives the customer markup, comparison, revision and approval steps.
The first deliverable may be prototype geometry only. A game-ready job can add high-poly work, optimized geometry, UVs, textures, normal-map baking, up to three LODs and optional rigging. Those stages are sequential and configurable, so 'geometry', 'complete asset' and 'rigged character' must not be treated as the same output.
Its practical value is extra production capacity with fewer vendor-management steps. The buyer still owns the acceptance bar: scale, topology, material setup, collision, skeleton, deformation, naming, performance and final behavior must be checked in the target DCC and engine.
HOW YOU USE IT
Trial, Indie and Pro are paid self-serve routes. The current plan page lists a $50 seven-day Trial that automatically renews to Indie, then $400/month Indie and $1,200/month Pro. API and custom integration access are Enterprise-only and require contacting sales.
Prototype geometry
Standard, High and Ultra prototype routes provide geometry for blockout, direction approval or further DCC work. Geometry alone does not imply UVs, textures, rigging, LODs or engine readiness.
O06O07O12Staged game-ready production
Jobs can move through high-poly, optimized geometry, UV, texture, rig, baking and LOD stages, with comments, free revisions before approval and explicit approval gates.
O08O09Enterprise integrations
The official Web API supports image-to-geometry submission, status or webhook handling and expiring result URLs. Unity, Blender, Unreal and Omniverse plugins plus custom automation are described separately.
O13O14O15O16INTERFACE & EXAMPLES
See how it works
Kaedim's quick-start guide shows an asset list with production stages and review progress.
Official source retrieved September 2026; interface version not specified.Source: Kaedim Documentation
REVIEW SCOPE
This review covers a bounded game-asset order from reference preparation through staged review and target-engine acceptance. It does not rank Kaedim against automatic generators or estimate studio-wide outsourcing replacement.
- Beginner friendlinessSome basics needed
- The browser submission, visual markup and staged revision flow lets a non-modeler begin from references instead of an empty mesh. Reliable briefs, orthographic views, topology, UV, PBR, rig, LOD, collision, performance and engine acceptance still require basic 3D judgment or a named reviewer, and the cash entry point is high for unguided learning.
- Agent integrationLimited integration
- An official Enterprise Web API exposes authenticated image-to-geometry submission, progress or webhook handling and result downloads, and official pages list DCC and engine plugins. Self-serve plans do not include the API, no official MCP was located, and public API documentation does not cover the service's complete artist-review, markup, revision and staged-approval loop.
Scope, integration requirements and sourcesChecked
Beginner scope: A first bounded static prop or environment module for someone who can prepare clean references and has access to basic 3D or engine review; not a release-ready hero character from one image.
Agent scope: Codex, Claude Code or another external orchestrator submits a bounded Enterprise image-to-geometry task, persists the result and starts a separate validation pipeline; not autonomous control of every human-assisted production stage.
Confidence: learning Medium · Agent High
- Kaedim Enterprise Web APIOfficial
- Access
- Enterprise access requires a devID, API key and refresh token, and may require a studioID. Keep credentials in a trusted server or local automation host; do not place them in a shipped game client. Unity, Blender, Unreal and Omniverse plugins are separate workflow integrations, not proof of a public self-serve agent API.
- Billing
- Trial, Indie and Pro publish cash prices and credit capacity, but API access is Enterprise-only with no public price. The calling agent, storage, webhook host, DCC automation, validation and failed outputs add separate costs; self-serve nominal unit prices cannot be projected onto the contract route.
- Getting results
- Send up to six authorized reference images with quality, polygon, size and project fields; preserve the request ID; poll boundedly or verify the webhook; download OBJ, FBX, GLB or GLTF before the result URL expires; persist files; then run deterministic DCC and engine checks and route failures to a human.
- Limitations
- Limited means a real but gated subset. Result URLs expire after one hour, authentication renewal and webhook security belong to the caller, the official example repository is thin, and the public API does not establish programmable access to artist review, all game-ready stages, revisions, approvals, rigging or LOD choices.
These are independent editorial judgments, not an overall score or a claim of hands-on agent integration testing.
- Kaedim Quick Start · Official · Learning evidence
- Kaedim input guidelines · Official · Learning evidence
- Kaedim game-ready workflow · Official · Learning evidence
- Trustpilot Kaedim user records · Independent · Learning evidence
- Product Hunt Kaedim user records · Independent · Learning evidence
- Kaedim Enterprise Web API · Official · Integration evidence
- Kaedim custom integrations · Official · Integration evidence
- Kaedim Unity plugin · Official · Integration evidence
- Kaedim API implementation example · Official · Integration evidence
- Kaedim plans · Official
- Kaedim company FAQ · Official
- Kaedim Privacy Policy · Official
KEY FINDINGS
What the evidence supports—and what it does not.
Official factHigh confidence
The human-assured workflow is the product
Kaedim now explicitly combines proprietary AI with internal artists or an Art Lead who checks and adjusts each delivery before customer approval. This explains both the production framing and why comparing only generation prices is misleading.
O01O03O04U09Official factHigh confidence
Prototype and game-ready outputs are different purchases
Prototype delivers geometry. Game-ready can add optimization, UVs, textures, rigging, baking and LODs through sequential stages. A result must be described by the stages actually included—not by a generic '3D model' label.
O06O07O08O09Editorial inferenceHigh confidence
Three public conflicts affect buying decisions
Current pages disagree on whether trial access is paid or free, whether Indie includes a game-ready asset, and whether the customer needs DCC polishing. We preserve all three conflicts and recommend checking checkout or obtaining written confirmation.
O02O03O06O11O18User reportsMedium confidence
Historical outcomes vary by input and object
Users reported high reference fidelity, low fidelity, useful topology, platform instability and different results across character and building tasks. Most detailed reports predate the current service positioning, so they establish risk themes rather than 2026 quality.
U01U02U03U04U05U06U07Official factHigh confidence
The API is real but does not expose the whole service
Enterprise clients can submit image-to-geometry jobs and retrieve files through status checks or webhooks. Public documentation does not establish programmable markup, artist review, all game-ready stages, revisions and approvals, and no official MCP was located.
O08O13O14O15O21Editorial inferenceMedium confidence
Commercial claims still need a current contract
The vendor says subscribed users own outputs, receive commercial rights and do not contribute project assets to shared model training. Yet a readable current Terms page was not located, the privacy policy is marked 2021, and human artists review assets; unreleased intellectual property needs project-specific terms.
O01O03O11O17EDITORIAL VERDICT
Our take
Medium confidenceConsider Kaedim when a defined asset is valuable enough to justify a high cash entry point and your team can specify and accept 3D work. The human-assured, approval-led service may be more useful than repeatedly drawing cheap candidates, but current independent evidence is too old and sparse to promise quality, speed or return on investment.
Conclusion scopeA small game team with concept art, an art bible, a technical asset budget and a named reviewer ordering a prop, environment module or bounded character. Not a conclusion about an unplanned beginner experiment, full scene, hero-character pipeline or bulk production contract.
BETTER FIT
Worth auditioning when
- Teams with clear concepts, multiple reference views, style rules and technical acceptance criteria.
- A valuable prop, module or character where managed revisions matter more than the lowest generation price.
- Studios that need additional 3D capacity without coordinating several freelance handoffs themselves.
- Enterprise teams willing to contract for API, automation, custom styles, security materials or an SLA.
POORER FIT
Do not depend on it yet when
- A beginner who expects one image to become a release-ready character without learning 3D acceptance.
- Low-cost, high-volume candidate exploration or casual learning below a $400 monthly budget.
- A project that cannot expose unreleased references to human reviewers without a negotiated contract.
- A coding agent that needs a self-serve API or MCP to automate the complete human review and approval loop.
WORKFLOW FIT
Where it fits in a production workflow.
Write the acceptance sheet first
Name the asset, target engine, scale, polygon range, topology, mesh split, UV and texture needs, rig template, LODs, collision, naming and file formats before ordering.
GuardrailIf nobody can judge these fields, appoint a 3D reviewer before paying for a complete asset.
Prepare one clean reference pack
Use one object on a clear background, add orthographic front, side, back, top and bottom views where possible, and separate complex accessories instead of hiding them in one silhouette.
GuardrailUpload only material you have the right and contractual clearance to share with human reviewers.
Choose prototype or named game-ready stages
Use prototype geometry for shape and direction. For a production asset, list the exact optimization, UV, texture, rig, bake and LOD stages required and confirm that the chosen plan includes them.
GuardrailDo not treat the word 'complete' as a substitute for written deliverables while public entitlement pages conflict.
Approve stages, then test the actual build
Annotate visible issues before approving each stage. Download the final files promptly, import them into the real DCC and engine, and check materials, collision, skeleton, deformation, frame budget and gameplay behavior.
GuardrailApproval is costly to reverse: keep screenshots, versions, comments and the final acceptance result for every asset.
RECOMMENDED WORKFLOW
Use this tool inside a complete production workflow.
Start from a testable brief, then move through prototyping, assets, audio, testing, and release with an explicit handoff and human check at every step.
SHORTEST RESPONSIBLE PATH
Start with one valuable static prop and a written acceptance bar
This is a low-risk starting workflow synthesized from the evidence.
- 01
Choose one static prop or environment module whose concept is already stable; avoid a hero character for the first order.
- 02
Prepare three to ten clean views and a one-page technical acceptance sheet for the target engine.
- 03
Recheck the live checkout, automatic renewal and exact stages included in Trial or Indie; save the terms shown at purchase.
- 04
Order prototype geometry first when shape or reference interpretation is uncertain, then decide whether an upgrade is justified.
- 05
Comment on each stage before approval, using the written acceptance sheet rather than general visual preference.
- 06
Import the result into the real project, measure integration and correction time, and divide total spend by assets that actually pass before scaling.
PRICING & RIGHTS
The public price buys plan capacity—not an accepted game asset
Pricing and terms last checked: Sep 12, 2026
- The current plan page lists a seven-day Trial at $50 for five credits and up to five geometries. It is marked 50% off and automatically renews to Indie unless cancelled.
- Indie is $400 per month for twenty credits, up to twenty geometries or one complete asset described there as including high-quality retopology, UVs and textures. The same row does not promise rigging or LODs.
- Pro is $1,200 per month for sixty credits, up to sixty geometries or three complete assets with the same listed retopology, UV and texture scope plus priority processing.
- Enterprise has no public price. API, custom integrations, styles and service commitments are contract scope, so self-serve arithmetic must not be projected onto an Enterprise order.
- Documentation assigns one, two or four credits to Standard, High or Ultra prototype geometry, while the plan page advertises maximum geometry counts without one reconciled entitlement table.
- Public pages conflict: the Games page says the first three assets are free while the documentation FAQ says there is no free trial; the plan page's $50 route is used only as this dated price snapshot.
COST TO OUTPUT
What each public dollar figure can nominally produce
Arithmetic from the maximum capacities on the September 12, 2026 plan page. It excludes unused capacity, quality tiers, Enterprise terms, input preparation, waiting, revisions, DCC/engine integration and acceptance failures.
O02O06O07O08O11O12O18That is a nominal $10 per geometry only when all five published capacity units are usable. The route automatically renews to Indie unless cancelled.
Nominally $20 per geometry or $400 per complete asset. Public documentation conflicts over whether Indie receives the full game-ready path.
The nominal unit prices remain $20 and $400. Priority processing does not establish delivery time, first-pass acceptance or engine readiness.
Indie needs ten monthly capacity blocks. Pro needs four and leaves room for twelve assets. This illustrates monthly rounding, not delivery concurrency, schedule or accepted quantity.
Current public evidence does not provide comparable acceptance rates, correction hours or exact entitlements, so a reliable cost per accepted geometry, complete asset, rigged character or scene cannot be calculated.
Actual accepted-asset cost = subscription or contract, reference preparation, revision and waiting time, DCC and engine integration, materials, rigging, collision, performance work and QA divided by assets that pass the project's bar. Keep that number unknown until your own fixed task is measured.
Commercial-use condition
Commercial use appears allowed with conditions. Kaedim says subscribed users own generated models, receive full commercial rights and do not contribute project assets to shared model training. However, a readable current Terms text was not located, input rights remain the uploader's responsibility, and exact Indie deliverables conflict across public pages. Save the purchase terms and obtain current contract language for important work. This is not legal advice.
Human artists or an Art Lead inspect and adjust assets. The visible privacy policy is marked April 14, 2021 and does not place the current asset workflow, personnel access, subprocessors, project-file retention and deletion terms in one complete statement. Treat unreleased characters, brands and art bibles as confidential and settle access, retention, deletion, security, refunds and automatic renewal in writing.
PRODUCTION RISKS
Resolve these before production use.
Plan and documentation claims conflict
Trial price, Indie game-ready access and customer cleanup have incompatible public descriptions. Check the live checkout and request written deliverables before paying; screenshots of a marketing page are not a contract.
O02O03O06O11O18A weak reference pack can waste an expensive slot
Single-view silhouettes, cluttered backgrounds, hidden sides and combined accessories create ambiguity. Validate references and prototype shape before committing a complete-asset slot.
O05O10U01U02U07Approval does not replace engine acceptance
A visually approved FBX or GLB can still fail scale, material, collision, skeleton, deformation, naming or performance requirements. Keep a target-build gate after the Kaedim workflow.
O08O09O11Current quality evidence is thin
Detailed user outcomes are mostly old; recent identifiable comments lack inputs, files and acceptance results. Do not infer a 2026 success rate, revision speed or comparison rank from vendor cases or aggregate ratings.
O19O20U01U02U03U04U10Enterprise automation still needs a wrapper and contract
Authentication renewal, webhook validation, one-hour result URLs, persistence, retries and quality gates belong to the caller. Public API pages do not expose the complete artist-review and approval loop.
O13O14O15O21NOT VERIFIED
Claims this page does not make
- Current first-pass acceptance rate, reference fidelity, topology quality, platform stability, revision turnaround and delivery schedule for representative game assets.
- Whether the live Trial is $50, the first three assets are free, or another checkout-specific offer applies to a particular account.
- The exact Indie complete-asset stages, including whether rigging, baking and LODs are available or require Pro or Enterprise.
- Whether each current delivery needs customer DCC polishing and how much engine-integration work remains.
- A readable current Terms text covering output rights, renewal, refunds, termination, liability and human access; current project-file retention and subprocessors.
- Cost per accepted prototype geometry, game-ready asset, rigged character or scene after plan capacity, failed work, integration and QA.
- A current public MCP, self-serve API or programmable end-to-end artist-review, revision and approval workflow for Codex or Claude Code.
- A defensible current comparison against Meshy, Tripo, Sloyd or traditional outsourced 3D production.
RESEARCH METHOD
Research-reviewed from public evidence
Research-reviewed from public evidence. We checked twenty-one official product, plan, workflow, API, integration, ownership and privacy sources, then coded ten principal external records across Product Hunt, Trustpilot, Reddit, independent reporting and an identifiable studio comment. Product Hunt and Trustpilot each contribute three of ten; no platform exceeds 30%. Positive, negative, mixed and disclosure evidence is retained, but most detailed outcomes are from 2022–2024 and vendor cases are not counted as independent quality evidence. No account, payment, upload, model generation or Enterprise API call was made or required.
Research windowCurrent official product, plan, documentation, API, ownership and privacy pages were checked on September 12, 2026. Ten principal external records span five platforms or carriers, but most detailed outcomes are from 2022–2024.
OFFICIAL SOURCES21
- O01Kaedim current productChecked Sep 12, 2026
Current positioning, ownership, security and shared-model-training claims.
- O02Kaedim plansChecked Sep 12, 2026
Current Trial, Indie and Pro prices, credits, capacity and automatic renewal.
- O03Kaedim company FAQChecked Sep 12, 2026
Artist-assured workflow, Enterprise scope, commercial claims and no-cleanup statement.
- O04Kaedim documentationChecked Sep 12, 2026
Documentation entry and human-in-the-loop product overview.
- O05Kaedim Quick StartChecked Sep 12, 2026
Asset submission, multi-view upload and review flow.
- O06Kaedim asset typesChecked Sep 12, 2026
Prototype, game-ready, rigged and scene deliverables, including the conflicting plan-access statement.
- O07Kaedim prototype assetsChecked Sep 12, 2026
Prototype geometry quality levels, settings, credits and processing estimates.
- O08Kaedim game-ready assetsChecked Sep 12, 2026
Game-ready settings, revisions, optional stages and processing estimates.
- O09Stages of a game-ready assetChecked Sep 12, 2026
The sequential high-poly, optimization, UV, texture, rig, bake and LOD workflow.
- O10Kaedim input guidelinesChecked Sep 12, 2026
Reference-image requirements, orthographic views and complex-object limits.
- O11Kaedim documentation FAQChecked Sep 12, 2026
Formats, intellectual property, polishing, revisions and the no-free-trial statement.
- O12Kaedim get startedChecked Sep 12, 2026
Credit schedule and Enterprise access context.
- O13Kaedim Enterprise Web APIChecked Sep 12, 2026
Enterprise-only authentication, generation request, status, webhook and result-download contract.
- O14Kaedim API overviewChecked Sep 12, 2026
API and user-generated-content integration overview.
- O15Kaedim custom integrationsChecked Sep 12, 2026
Plugins, scripts, automation and custom Enterprise integration scope.
- O16Kaedim Unity pluginChecked Sep 12, 2026
Documented Unity plugin handoff and import workflow.
- O17Kaedim Privacy PolicyChecked Sep 12, 2026
Privacy policy marked last updated April 14, 2021.
- O18Kaedim for gamesChecked Sep 12, 2026
Game-team workflow and the conflicting first-three-assets-free statement.
- O19Kaedim release logChecked Sep 12, 2026
Public documentation release history.
- O20Little Buffalo Studios case studyChecked Sep 12, 2026
Vendor-selected game-studio workflow and savings claims; not independent quality evidence.
- O21Kaedim API implementation exampleChecked Sep 12, 2026
Official public API example with a thin repository history and no releases.
INDEPENDENT EVIDENCE RECORDS10
- U01Product Hunt review by WesChecked Sep 12, 2026
A simple-character attempt reported a long wait, low reference fidelity and unfinished fingers.
- U02Product Hunt review by madebyjoe studioChecked Sep 12, 2026
A detailed multi-view character was reported near the reference and expensive but worthwhile.
- U03Product Hunt review by Amy JiangChecked Sep 12, 2026
A historical report of upload and generation instability.
- U04Trustpilot review by Alice RobertsonChecked Sep 12, 2026
A self-described 2D designer reported that submission and revision tools helped them begin 3D work.
- U05Trustpilot review by Karl-Johan StrandbergChecked Sep 12, 2026
A promotional-sounding one-review account reported useful topology, detail and hours saved; weighted low.
- U06Trustpilot review by EmChecked Sep 12, 2026
A basic-sketch character attempt reported satisfaction without downstream details.
- U07Reddit mixed-quality three-run accountChecked Sep 12, 2026
One user reported good character topology but poor results for a Paris building facade across three runs.
- U08Reddit trial and payment-friction accountChecked Sep 12, 2026
A user expected a free trial but encountered a payment request; unsupported accusations are excluded.
- U09TechSpot report on historical human-labor disclosureChecked Sep 12, 2026
Independent reporting on earlier human involvement and disclosure; current pages now state artist assurance.
- U10Redly Games customer commentChecked Sep 12, 2026
A recent identifiable studio comment praised the collaboration but supplied no task or acceptance detail.
DISCLOSURE
How this research was supported
Independent editorial research. MakeGameWithAI has no affiliate link, sponsorship, vendor-provided account, free credits, asset generation, interview or technical support for this review. Owner Review approved the managed-service positioning, some_basics / medium beginner decision, limited / high external-Agent decision, nominal cost framing, commercial and data boundaries, and three-conflict disclosure on September 12, 2026, then approved publication on September 14, 2026.
This public research review will be revisited when pricing, terms, product versions, or material new evidence changes.