Sloyd Review 2026

Rodin (Hyper3D)

Photorealistic AI 3D models with 4K PBR textures, generated from a single photo

Rodin, built by Deemos and marketed under the Hyper3D brand, sits at the premium end of AI 3D generation. Where budget tools produce game-ready but visibly "AI-generated" meshes, Rodin specifically targets photorealistic output — quad-mesh models with 4K PBR (physically-based rendering) textures generated from a single image or multiple reference photos, aimed at product visualization, architecture and high-end asset work rather than quick game props. Its credit-based pricing reflects that positioning: previews are free to generate, and credits are only spent when you actually download a model, letting you iterate without burning budget on rejected attempts. Compare that to faster, cheaper options like [Meshy](/tools/meshy) or [Tripo AI](/tools/tripo-ai) in our [AI 3D & Avatars category](/category/ai-3d) — those are the better choice for game-ready assets at volume, while Rodin is the one to reach for when visual fidelity genuinely matters more than speed or cost per model. An 8x-faster "Speedy" mode trades some fidelity for quick iteration, and both single- and multi-view image input are supported, giving more control over the final geometry than text-prompt-only competitors.

Meshy

Fast text-to-3D and image-to-3D generation with free rigging and animation

Meshy generates textured, riggable 3D models from text prompts or reference images in seconds, aimed at game developers, 3D-printing hobbyists, and AR/VR creators who need fast first-draft assets rather than production-clean topology out of the box. Now on Meshy 6, it supports text-to-3D, image-to-3D, and text-to-texture generation across multiple art styles (realistic, cartoon, low-poly, voxel), with automatic PBR texture maps (metallic, roughness, normal) for realistic engine lighting. A genuinely generous feature: auto-rigging and animation currently cost 0 credits — unusually free for the category — with 500+ built-in animation presets (walking, running, combat moves) that make generated characters immediately usable in Unity or Unreal Engine. For the 3D-printing community specifically, its Printability Check and one-click Bambu Studio slicer integration report a 97% slicer pass rate, meaningfully reducing failed prints from geometrically invalid meshes. Pricing is straightforwardly credit-based: Free gives 100 credits/month under a CC BY 4.0 license, Pro is $20/month for 1,000 credits, and Studio is $60/month for 4,000 credits, both including API access and private assets. A complete generation (mesh plus texture) costs 20 credits regardless of plan, so cost-per-model is identical across tiers — you're paying for volume and privacy, not cheaper individual generations. The honest limitation, consistent across reviews: output quality varies, works best for organic shapes and rapid prototyping, and commercial-quality characters or precise hard-surface assets for professional game/film pipelines often need manual cleanup afterward.

Tripo AI

Trending

Fast text- and image-to-3D model generation

Tripo AI converts a text prompt or a single reference photo into a textured 3D mesh in well under a minute, which is a different workflow from traditional modelling: you iterate on concepts at sketch speed and only refine the winner. Outputs come with PBR textures and can be auto-rigged for humanoid shapes, and the retopology option produces cleaner quad-dominant geometry suitable for animation rather than the dense triangle soup early text-to-3D tools produced. Exports cover GLB, FBX, OBJ, USDZ and STL, so assets drop straight into Unity, Unreal, Blender or a 3D printer slicer. Game studios use it for background props and prototype assets, ecommerce teams for quick product visualisations, and AR developers for lightweight models that do not need a full art pipeline.

Spline AI

Text-to-3D generation inside a full interactive design and publishing platform

Spline AI is the AI 3D Generation feature built into Spline, a browser-based 3D design tool — rather than a standalone text-to-3D generator, it's positioned inside a broader platform with an editor, publishing pipeline, and multi-surface export (web, Apple, Android). That platform context is the real differentiator: the value isn't just generating a 3D object quickly, it's being able to keep working with that object inside a usable, exportable 3D toolchain afterward, rather than downloading a raw mesh and starting over in different software. As of its April 2026 "AI 3D Generation V1" release, Spline AI supports text-to-3D, image-to-3D from front-facing 2D reference images, and variant remixing — generating multiple 3D options from one prompt to choose from. It's explicitly positioned as beginner-friendly, designed to let people create 3D assets without traditional modeling software knowledge, while still producing objects that plug directly into Spline's interactive scene-building and web-embedding features. Access requires an active paid Spline plan plus the separate Spline AI Add-on, with generation metered by AI credits — it's not available on a free tier. Starting around $9/month for entry-level Spline access before the AI add-on, it's best suited to designers and creators who want 3D generation as part of a broader interactive web/app design workflow, not people who just need a standalone 3D asset exported for a completely separate pipeline like a game engine — for that, a dedicated tool like Meshy is generally a better fit.

Ready Player Me

Cross-platform avatars from a selfie, usable in 600+ apps and games

Ready Player Me (originally launched as Wolf3D) solves a specific interoperability problem: instead of creating a separate avatar for every game or virtual world you use, it generates one 3D avatar from a selfie that works across 600+ compatible apps and games — VRChat, Spatial, and other metaverse platforms — carrying your digital identity between experiences rather than starting fresh each time. Trusted by major names including VRChat, Ubisoft, Spatial and Adidas, it's become a genuine cross-platform standard rather than a single-game novelty. Its AI-driven pipeline (RESTYLE, Asset Morphing, Shape 3D) automates styling, rigging and optimization so avatars look consistent and correctly proportioned across different game engines and device capabilities, from mobile to VR headsets. Any developer can integrate it into their own apps and games using a free avatar SDK, compatible with Unity and Unreal Engine and working across web, mobile and desktop — that free developer access is a core part of why it's spread across so many compatible experiences. With over a decade of avatar R&D behind the underlying hardware-to-software transition and a community of 25,000+ integrating developers, it's positioned as infrastructure for the broader interoperable-avatar concept rather than a single consumer product with one clear price tag — enterprise and custom integration pricing is generally negotiated directly rather than published in flat tiers. For developers building any experience where a portable, cross-game avatar identity matters, the free SDK access makes it a low-friction default; for end users, it's simply the avatar creation layer sitting behind many of the virtual worlds they may already use without necessarily realizing it.

Luma Genie

Fast 3D asset generation, now bundled into Luma's Agents subscription

Luma Genie generates textured, game-ready 3D assets from text prompts in under three minutes, with automatic PBR map generation (normals, roughness) for realistic engine lighting and native UV unwrapping, exporting directly into formats compatible with Blender, Unity and Unreal Engine. Its low-poly, optimized topology is specifically well-suited to mobile and web AR applications where file size and rendering performance matter more than maximum surface detail. Like Luma's Dream Machine video product (covered separately in AI video tools), Genie is now bundled into the broader "Luma Agents" platform rather than sold as a standalone subscription — as of the March 2026 relaunch, there is no standing free tier to test it independently; access requires a Luma Agents plan (Plus $30/month, Pro $90/month, Ultra $300/month), with Genie included as part of the toolset rather than metered separately. This mirrors the same pricing restructure affecting Luma's video generation, worth knowing if you were expecting an older, cheaper standalone Genie subscription. The honest limitation, consistent across reviews: extremely fine organic details (human eyes, expressive facial features) can appear soft, and Genie lacks native in-browser mesh-editing tools for localized geometric adjustments — you'll typically export to Blender or similar for fine cleanup. For teams already invested in Luma's broader Agents platform for video generation who want 3D asset creation bundled into the same subscription, Genie is a convenient inclusion; for someone who only needs 3D generation specifically, a dedicated tool like Meshy at a much lower entry price is likely the more cost-effective choice.

Kaedim

AI 3D generation with human artist quality review before delivery

Kaedim takes a fundamentally different approach from fully-automated 3D generators: upload a concept art image, and its AI generates a 3D model that a human artist then reviews and cleans before delivery — combining generation speed with human quality assurance. The result is consistently cleaner than fully-automated alternatives, with better topology, proper UV layouts and more accurate proportions, at the cost of turnaround measured in hours rather than the seconds or minutes competitors like Meshy offer. Its core limitation is also its defining characteristic: Kaedim requires image input specifically and doesn't support text-to-3D generation — you need a concept art reference to start from, positioning it squarely for studios and creators converting existing concept art libraries into 3D at scale, rather than ideating from scratch. Batch processing large libraries of concept art is a genuine strength for production pipelines needing to convert many assets consistently. Pricing is per-asset with volume discounts and enterprise plans rather than a flat monthly subscription, reflecting its human-in-the-loop production model. It's the highest-quality output among automated-plus-human-review 3D tools, but also the most expensive per asset and slowest to deliver — the right choice specifically when clean, animation-ready topology matters more than speed or cost, such as production game or film pipelines; for rapid prototyping or hobbyist use where iteration speed matters more than polish, Meshy or Luma Genie's fully-automated, near-instant generation is the more practical fit.

CSM (Common Sense Machines)

Fast image-to-3D generation now owned by Google, after its January 2026 acquisition by Alphabet

CSM (Common Sense Machines) converts a single 2D image, a short video, or a text prompt into a textured 3D mesh in under two minutes — genuinely useful for rapid prototyping of background props and environment pieces, though the resulting topology is rougher than purpose-built competitors and typically needs manual retopology for anything going into a rigged, animated asset. An important update worth knowing before you evaluate it: CSM was acquired by Alphabet (Google) in January 2026, with its team joining Google DeepMind. The product remains operational with its existing pricing tiers as of this writing, but its long-term roadmap as a standalone tool is genuinely uncertain post-acquisition — worth checking CSM's own site for the latest status before committing to a paid plan. Compared to [Rodin (Hyper3D)](/tools/rodin-hyper3d) in our [AI 3D & Avatars category](/category/ai-3d), CSM trades geometry quality for speed and a lower entry price, with plans from a free Tinkerer tier (10 one-time credits, CC BY 4.0 licensed output) up to $60/month for private, commercially-licensed generations. It's best suited to indie developers and concept artists who need a fast 3D starting point, not final production-ready assets.

3D AI Studio

Generation, texturing and auto-remeshing in one no-code tool — the well-rounded pick in AI 3D

3D AI Studio positions itself as the generalist option in AI 3D generation: rather than specializing narrowly like [Sloyd](/tools/sloyd)'s parametric game props or [Rodin](/tools/rodin-hyper3d)'s photorealistic focus in our [AI 3D & Avatars category](/category/ai-3d), it combines image-to-3D and text-to-3D generation with built-in AI texturing and auto-remeshing in a single no-code interface — handling more of the full pipeline in one tool rather than requiring you to export to separate software for cleanup. That auto-remeshing step matters more than it sounds: raw generative 3D output is frequently unusable for animation or clean game-engine import without retopology, and having that step built in rather than as a separate paid tool or manual Blender pass saves real production time. Supported export formats cover the standard game-engine and 3D-software needs (FBX, OBJ, GLB), keeping it compatible with existing Unity, Unreal or Blender workflows. A free tier lets you test generation and texturing quality before committing; paid plans start around $9.99/month, positioning it as one of the more affordable full-pipeline 3D tools in the category rather than the cheapest single-purpose option or the most premium quality leader.

Wonder Dynamics (Autodesk)

AI that auto-animates and lights a 3D CGI character onto live-action footage, now backed by Autodesk

Wonder Dynamics solves a problem that used to require a VFX studio: turning a live-action video into a scene with a fully animated 3D CGI character, automatically. Upload footage of an actor's performance, choose or upload a 3D character, and its AI analyzes the actor's motion, matches the scene's camera movement and lighting, and automatically animates, lights and composites the character into the shot — a process that traditionally needed motion-capture equipment and manual VFX compositing work. The company was acquired by Autodesk in 2024, and the tool continues operating with ongoing development backed by Autodesk's broader creative-software ecosystem — a meaningfully different stability profile than an independent startup in this category, worth knowing since several AI 3D tools in [our category](/category/ai-3d) have shut down or pivoted in the past year. Compared to [Rodin (Hyper3D)](/tools/rodin-hyper3d) in the same category, Wonder Dynamics isn't a 3D model generator at all — it's an animation and compositing pipeline for existing 3D characters onto live footage, aimed at indie filmmakers and small studios who can't afford a traditional VFX pipeline. Access and pricing are managed through Autodesk's platform rather than a simple self-serve subscription; check Autodesk's current offering for exact terms.

Tripo AI

Trending

Fast text- and image-to-3D model generation

Tripo AI converts a text prompt or a single reference photo into a textured 3D mesh in well under a minute, which is a different workflow from traditional modelling: you iterate on concepts at sketch speed and only refine the winner. Outputs come with PBR textures and can be auto-rigged for humanoid shapes, and the retopology option produces cleaner quad-dominant geometry suitable for animation rather than the dense triangle soup early text-to-3D tools produced. Exports cover GLB, FBX, OBJ, USDZ and STL, so assets drop straight into Unity, Unreal, Blender or a 3D printer slicer. Game studios use it for background props and prototype assets, ecommerce teams for quick product visualisations, and AR developers for lightweight models that do not need a full art pipeline.