MakeUseOf framed the sale as confirmation that Thingiverse had become a secondary destination for serious makers. The evidence supports a narrower conclusion. Thingiverse’s value is now largely historical: it is still useful when hunting an old bracket, replacement part, enclosure, or remix that has not migrated elsewhere. For a model that needs to print correctly on the first attempt, the better starting point increasingly depends on the printer, slicer, and type of object—not on Thingiverse’s once-unmatched name recognition.
MyMiniFactory’s purchase is real, and it is consequential. But it is a rescue plan, not proof that the rescue has worked.
MyMiniFactory bought an archive, not a solved platform
MyMiniFactory announced that it had acquired 100% of Thingiverse on February 12, describing the deal as a combination of Thingiverse’s huge legacy community with its own creator-first marketplace. MakerBot later acknowledged the transfer publicly. Tom’s Hardware and TechRadar also reported the acquisition and the new owner’s intention to maintain Thingiverse as its own platform while bringing it into the SoulCrafted umbrella.
The central promise is a hard one: MyMiniFactory says it intends to reduce and eventually remove AI-generated designs from Thingiverse. That responds to a visible problem across model repositories, where low-effort uploads, copied listings, misleading preview images, and untested geometry can overwhelm useful search results.
But MyMiniFactory has not publicly supplied the details that determine whether that promise can improve daily use. There is no published, platform-specific explanation of how an upload will be identified as AI-generated, whether older uploads are being reviewed, what happens to a model falsely removed, or whether the company will distinguish between wholly generated geometry and a designer using AI-assisted tools during modeling. “Eventually remove” is a direction of travel, not a moderation system.
There is also a notable mismatch in the acquisition’s headline statistics. MyMiniFactory’s announcement referred to a design library of more than six million files, while contemporary reporting on the deal described Thingiverse as holding about 2.5 million “things,” alongside nearly eight million users. Those figures may be counting different objects—files, designs, revisions, collections, or users—but neither company has explained the difference. For makers, the distinction matters: a larger raw count does not tell you how many listings remain downloadable, printable, attributed, or maintained.
The acquisition does not yet fix search quality, file provenance, or print validation. Those are the practical reasons many users stopped making Thingiverse their first tab.
Thingiverse still has a role as the long tail
Writing Thingiverse off would be a mistake. Its age is its advantage.
MakerBot launched Thingiverse in 2008, and for much of the following decade it served as the default public archive for desktop 3D printing. That history created an unusually deep catalog of printer modifications, Raspberry Pi cases, electronics enclosures, workshop fixtures, replacement knobs, vintage-project parts, and remixes that may not exist in the same form elsewhere.
For Windows hobbyists, that makes Thingiverse a useful second-pass tool. If a search on Printables, MakerWorld, or a manufacturer’s GitHub repository turns up nothing, Thingiverse may be where an older solution survives. This is especially true for discontinued printers and electronics projects whose communities peaked before newer repositories became dominant.
The correct workflow is to treat it as an archive rather than a recommendation engine. Do not assume a high download count means a model is current, dimensionally correct, compatible with a modern printer, or licensed for the intended use. Read the description, check the upload date and comments, inspect makes from other users, and open the STL or 3MF in your slicer before committing filament.
A model that has a clear bill of materials, photographs of the assembled object, documented hardware sizes, orientation advice, and user print reports is more valuable than a polished thumbnail. That has always been true, but it becomes essential when an archive has years of abandoned uploads mixed with current work.
Printables is the stronger general-purpose starting point
Prusa’s Printables has become a sensible default for users who care more about repeatable prints than catalog size. It is open to non-Prusa owners, and its contests, collections, community makes, and print-profile culture make it easier to judge whether an object has been printed outside its creator’s CAD viewport.
The platform’s biggest practical advantage is its connection to the desktop workflow. Prusa documents direct Printables integration in PrusaSlicer, including one-click import and an in-app Printables tab. That does not mean every uploaded model is trustworthy, or that Prusa’s recommended profiles will be right for every machine and filament. It does mean the site is designed around the handoff from browser to slicer rather than treating the STL download as the end of the product.
MakeUseOf also described Printables as offering direct plug-in integration with both PrusaSlicer and OrcaSlicer. The Prusa documentation supports the PrusaSlicer portion of that claim, but it does not establish an equivalent official one-click OrcaSlicer integration. OrcaSlicer users can still download standard STL and 3MF files from Printables and import them normally, but that is different from a documented native workflow.
For a Windows user running a mixed fleet—perhaps a Prusa machine at one desk and a Bambu Lab, Creality, or custom Klipper printer elsewhere—Printables has another advantage: its files are generally not tied to one vendor’s machine-specific print profiles. That makes it a better first choice for functional parts where material, nozzle, wall count, and tolerances need to be chosen locally.
MakerWorld is fast, but its incentives shape the catalog
MakerWorld is the most convenient choice for Bambu Lab owners because it is closely wired into Bambu Studio and the company’s printers. A user can find a model, select a community or creator profile, and send a job into the familiar Bambu workflow with little friction. That convenience is real, particularly for a P1S, X1-series, or A1 owner who wants a quick household print rather than a CAD project.
Bambu Lab’s own posts also reveal why convenience should not be confused with curation. The company has repeatedly changed MakerWorld’s reward system after identifying manipulated downloads, print trading, fake engagement, and incentives that favored simple, high-volume uploads. It has also temporarily disabled sharing features in the past after saying that complaints over shared models greatly exceeded the share function’s small proportion of the catalog.
Those disclosures matter because rewards influence what creators upload and what users see. A platform that pays for downloads, boosts, print profiles, or exclusivity will attract high-quality designers, but it also creates reasons to optimize for visibility rather than durability or originality. MakerWorld has moderation and verification mechanisms, yet its growth makes it vulnerable to the same discovery problem facing every open-upload service.
Use MakerWorld first when you have a Bambu printer and a model has substantial evidence behind it: multiple completed makes, photographs that show the actual object rather than only renders, a suitable profile for the intended material, and comments confirming fit. For a load-bearing mount, threaded component, replacement part, or object involving heat, do the same independent checks you would on any other repository.
The site is less important than the evidence attached to a model
Paid marketplaces such as Cults3D and MyMiniFactory can be useful when a designer’s support, documentation, and production-ready detail justify the price. A paid file is not automatically better engineered, and free files are not automatically abandoned. The meaningful difference is accountability: designers selling a model under their own name have stronger incentives to update instructions, answer questions, and protect their reputation.
Thangs serves a different purpose. Its strength is discovery across repositories, which helps when the requirement is precise: a DIN-rail mount for a particular mini PC, an adapter for a named tool, a rack bracket with a fixed hole pattern, or a replacement component whose geometry matters more than appearance. Search results still need to be evaluated at the original host, because the file’s license, documentation, version history, and community feedback travel poorly through a search layer.
The most reliable repository choice therefore starts with the job:
- Use Printables when you want broadly compatible files backed by community makes and a straightforward desktop-slicer workflow.
- Use MakerWorld when a Bambu-specific profile and rapid send-to-printer convenience outweigh the risks of a heavily incentivized upload platform.
- Use Thingiverse when you need its legacy inventory, particularly older designs or remixes that newer repositories do not carry.
- Use a paid designer marketplace when instructions, support, and proven production quality matter enough to justify the cost.
- Use cross-platform search when the part is specific and the first repository search fails.
Thingiverse’s new owner has identified the right problem: repository quality collapses when users cannot tell tested work from generated, copied, or abandoned uploads. The company now needs to show the unglamorous operational work behind the slogan—better search, usable reporting tools, transparent moderation, preserved attribution, and a review path for creators.
Until then, Thingiverse is best treated as a deep legacy vault. It may still contain the exact model a Windows-and-3D-printing setup needs, but it should no longer be assumed to be the place where that search begins.