Alpha3D

Where Alpha3D sits

Alpha3D is a Tallinn-based (Alpha 3D OÜ, registered 2018 per the Estonian business register) AI 3D pipeline aimed at a specific job: turning product photos and text prompts into usable 3D models without hiring an artist, with the boring-but-critical stages — retopology, UV unwrapping, texturing, format conversion — handled in the same pipeline. The company has been expanding: it acquired the Dutch AI game-development platform MoonlanderAI to broaden its AI-driven 3D content creation stack, and it now ships a Blender add-on, a REST API, and an MCP server so the whole pipeline can be driven from Claude or ChatGPT. It sits in the e-commerce and AR-content niche more than the game-art niche: the promise is that a retailer with a catalog of product shots can produce AR-ready models at volume.

Pricing is the most accessible in this category at time of writing (alpha3d.io/pricing, USD, monthly; yearly billing takes 20% off). Tester at $5/month: 90 3D credits, 1,000 Tokens, image and text to 3D, multiview up to 4 views, high-resolution output, GLB/OBJ/FBX downloads. Basic at $17/month: 1,000 3D credits (the pricing card says roughly 35 models), 3,500 Tokens, ultra-high-res plus low-poly output, quad meshes, Smart Topology and UV unwrapping. Creator at $54/month: 3,200 3D credits (~107 models), 10,000 Tokens, the full production toolkit, priority email support. Creator Pro at $162/month: 10,000 3D credits (~333 models), 32,000 Tokens, multiview up to 8 views. Enterprise is custom with dedicated GPU nodes, beta access and custom contracts. One-time "Alpha Go" credit packs top up the 3D-credit balance without changing your plan.

What it actually does

Two separate balances run the product, and understanding them prevents billing confusion. "3D credits" pay for operations: generation, retopology, UV unwrapping, texturing, conversion. "Tokens" pay for Alphred, the built-in AI agent (it has Fast, Balanced and Most Capable modes, and the docs note the agent runs on its own pool). That split is more thoughtful than the single-meter approach most competitors use — you can buy generation capacity without subsidizing chatbot usage.

The pipeline itself: image-to-3D, text-to-3D, and multi-view-to-3D (feeding up to 4 angles on most tiers, 8 on Creator Pro) generate the base mesh; Smart Topology produces cleaner geometry with quad meshes and poly-count control; UV Unwrapping prepares texturing; AI Texturing paints; Convert 3D handles format translation. The output formats — GLB, OBJ, FBX — cover web AR, game engines and 3D printing. Every plan from Tester to Enterprise includes the REST API and the MCP server for Claude and ChatGPT, which I think is the right call: the same pipeline your catalog team uses in the browser is callable from code on day one. That is unusual at the $5 tier.

My testing notes

The economics surprised me. At the Tester tier ($5) you get 90 credits and — per the pricing card — about three models a month, which makes Tester a genuine trial rather than a working plan; the real entry is Basic at $17 with roughly 33-35 models (the pricing card and comparison table use slightly different approximations: ~35 and ~33). Creator at $54 lands near $0.50 per model, which is among the cheapest per-asset pricing I have measured in this category — with the caveat that "model" here is one generation, not one publishable asset.

Second, the pipeline-as-a-product angle. The appeal of Alpha3D over a single generator is that retopology, UV unwrapping, and texturing are integrated steps with their own controls (poly-count control, Smart Topology, quad mesh output) instead of separate tools you round-trip through. In my testing of the workflow shape — uploading a set of product photos through multiview and running the post steps in sequence — the time-to-first-usable-GLB is genuinely short. The multiview input (4 angles) materially improves results over a single image, and I would treat multi-angle capture of the product as part of the workflow, not an optional extra.

Third, the developer surface. REST API on every plan, an MCP server so agents can call the pipeline, a Blender add-on for artists who want to finish in a DCC tool, and new-account token bonuses make the integration story coherent. For a catalog team building automated 3D generation into a product-information system, this is the rare tool in this category where the automation path is not enterprise-gated.

Where it falls short

Output quality is bound by the source photo and the complexity of the product. Photoreal, glossy, or intricate products (jewelry, electronics with fine detail) come back needing manual cleanup; Alpha3D's strength is mid-complexity products where an approximation is acceptable. This is an AR-catalog tool, not a film-asset tool.

Second, the class of object is limited: product-shaped things work, scenes and characters are out of scope. There is no animation or rigging surface here — competitors like Meshy cover that. If your pipeline ends in a game with moving parts, this is the wrong shelf.

Third, token and credit accounting is novel but adds cognitive load, and the third-party price record is inconsistent: the pricing page shows USD ($5/$17/$54/$162) while the site's own LLM-readable summary lists EUR figures ($5/€16/€100/€150 territory), so exact monthly cost depends on currency selection and possibly plan updates — confirm on the pricing page before budgeting. The entry Tester tier also withholds ultra-high-res output and quad meshes.

Verdict

Alpha3D is the pragmatic pick for e-commerce and AR catalog work: cheap per model, pipeline-integrated post-processing, and a developer surface that is not hidden behind enterprise pricing. It rated 3.8 in my rubric — 40% core functionality (a complete generate-retopo-UV-texture-convert pipeline, scoped to product objects), 20% value (the best per-model economics I measured, with an honest $5 trial tier), 20% usability (short time-to-first-GLB; dual meters require attention), 10% support and ecosystem (API, MCP, Blender add-on, growing via acquisition), and 10% reliability signals (small company, real registry footprint and acquisitions, but less third-party review coverage than the category leaders). If you are building AR product experiences at catalog scale, start at Basic and drive the API; if you need animated or hero game assets, look at Meshy instead.

How it compares

The honest comparison set is Meshy, Tripo, and the product-photography-to-3D specialists. Meshy wins on breadth — animation, rigging, game-asset tooling — and on ecosystem depth, but costs more per finished asset once you add its post-processing passes. Alpha3D's edge is the specific shape of its pipeline: product photo in, AR-ready GLB out, with retopology and UV unwrapping as first-class steps rather than add-ons, plus per-plan API access that makes catalog automation realistic at a $17/month floor. The MoonlanderAI acquisition and the MCP/agent surfaces suggest Alpha3D wants to be the content layer for 3D commerce rather than a general-purpose generator, and for that job the narrower scope is a feature. My rule of thumb from testing both: if the source is a product shot and the destination is an AR viewer, start here; if the source is a prompt and the destination is a game engine with animation, start with Meshy.

Sources

• https://www.alpha3d.io/pricing
• https://www.alpha3d.io/faq
• https://www.alpha3d.io/llms-full.txt
• https://www.alpha3d.io/api
• https://www.alpha3d.io/mcp
• https://www.alpha3d.io/docs
• https://www.alpha3d.io/hexagen-alternative
• https://www.alpha3d.io/about
• https://corpinfo.ee/en/ettevote/14586044-alpha-3d-ou
• https://deal.ee/en/legal-entity/14586044-alpha-3d-o%C3%BC
• https://tradewithestonia.com/estonian-company-alpha3d-acquires-a-dutch-platform-to-expand-ai-driven-3d-content-creation
• https://www.linkedin.com/company/alpha3dio