You can quickly build multiple models from text prompts or image links via the API interface and have the pipeline use the download links to retrieve the models. The API interface was clearly documented and easy to implement. Even for somebody who was a bit rusty with the latest coding styles. Review collected by and hosted on G2.com.

Some of the models 10-15% like all A.I are not perfect. Extra arms, difficult poses to print. The odd floating object, which you can't 3d print. But you can fix the models in in 3d tools way faster than creating a whole model from scratch. Even saying this the support team have refunded credits for some of these. Some of my prompts could have been better. Review collected by and hosted on G2.com.

Super easy to create high quality 3-D Models and interface is easy to navigate. Another great feature is the ability to use an API to connect with my coding agent. Performance is actually impressive, and I usually wait no more than two or three minutes per model. Pricing is quite reasonable as well with a typical 10-20 credits per model and joining and being given so many promotional credit opportunities for simple methods is really nice, like joining Discord or favoriting 3d model artists in their community. Of course, the AI agent does a good job at understanding both simple and complex prompts. I've not needed support yet but they do have an expansive knowledge base. Review collected by and hosted on G2.com.

That the rigging doesn't quite work well for non-human with tails or digitigrade legs. Review collected by and hosted on G2.com.

What I like best about Meshy is how quickly it helps turn an idea into a usable 3D asset. It makes 3D creation more accessible, especially for beginners, while still leaving room for creative decisions, editing, optimization, and integration into tools like Blender or Unreal Engine. Review collected by and hosted on G2.com.

It is a bit expensive for education. I think Meshy should offer free or special plans for schools, teachers, and students, because it has huge potential as a classroom tool. Review collected by and hosted on G2.com.

Right now, the image-to-3D model feature is kind of garbage for anything anime- or cartoon-like. That said, the chibi figure feature is way better. Even though it doesn’t produce super high-quality faces (since it’s chibi style), the rest of the model is usually pretty spot-on.

I don’t normally print the truly chibi proportions (short bodies with big heads), but I really enjoy the ones that keep the body and head more proportional. With those, it consistently creates a model I’d actually 3D print—except when the uploaded character has parts hidden behind their body. For example, if a hand or a sword is behind the character and out of view, the AI doesn’t seem to know how to complete it as a full object. Hands end up messy, and swords can come out super bent, fused into the character, or with the hilt starting on one side of the character and only the tip of the sheath showing on the other side.

I’d love to see something like the chibi figure approach applied to the image-to-3D model feature, so it could produce truly high-detail models that are ready to 3D print.

I’m also doing this just for fun and personal use with my own created 3D models. If there were a more affordable payment plan for solo creators, I think a lot more 3D model makers would flock to Meshy. Alternatively, either reduce the price (credit cost) for making models and/or give more credits through the daily AI model rating system.

For example, about a week ago, generating chibi figures cost 24 credits for 4 models to choose from. After picking one, it cost 20 credits to turn it into a full figure to print. Now it costs 30 credits to print, which means that if you earn 200 credits by doing the daily model rating, over the span of two days you end up making 2–3 fewer models.

Meshy, please don’t keep increasing the credit cost for making models, especially when the uploaded image doesn’t always give the results we want. It makes the end user feel like they wasted 24 credits on a model (or even more trying to convert it into a 3D-printable one), only for the final output to not be what we wanted or to look messed up from different angles.

Performance wise Meshy is fast, but could use some improvement on the end product. If I could choose a slower option to have a higher quality model in the end, I would choose the slow one. I don't know much about support because I haven't had an issue that needed reporting to admins.

Overall, though, Meshy is a good product, and I hope it keeps improving over time. Review collected by and hosted on G2.com.

The image-to-3D model generator is less than desirable. Out of 10 different images across various kinds of art and styles, I only end up liking about 1/5 of the results. It also struggles with faces at times, and with objects that aren’t clearly visible in the source image. For example, hands placed behind the head or weapons that aren’t fully in view often come out warped or just really messed up.

Honestly, the chibi figure program they have is superior to the image-to-3D model generator. That said, they recently increased the price to generate a chibi figure from 20 credits to 30 credits, and you definitely feel that over long-term use of the program.

One improvement I’d like to see in the chibi figurine generator is the ability to save the four generated images used as the base for making a figure, especially if I like more than one of the options. Right now, you can only use one of the four images, and the others are basically lost.

They should also let the user retry the finished model up to four times, like they do in the image-to-3D model generator, because sometimes the final product still has warped hands, incomplete parts, or messed-up weapons (usually the parts that weren’t visible in the original image, as I mentioned earlier). As an end user, I never want to feel like I wasted credits because what’s shown in the image doesn’t match what the AI produces at the end.

I understand AI is evolving every day and getting better, but a user shouldn’t feel like they’re being punished for only getting one shot to make things right. A few times, the final product had too many fingers, or the sword was just the hilt on the hip with the scabbard melded into the other side of the hip, and I thought to myself, “welp, that’s 54 credits I wasted on something not worth printing out.” If I had four retries, or even a way to retry and tell the AI what to remove for, say, 10 credits, I’d feel more satisfied—even if I still didn’t get the exact end product I wanted. Review collected by and hosted on G2.com.

I initially tried Meshy out on a whim because it seemed interesting, and I had seen some ads around so figured I can try it out for a month. That was 8 months ago and I'm still actively using it. It's kind of cheesy to rate it a "10" on the how likely am I to recommend it to others, but I've already recommended it to several and showed how awesome it can be in a workflow, so...I guess that has to be a 10, right?

The UI is clean, and easy to navigate, and it keep seeing updates which is nice.

As for the integrations, I've been using it primarily with Blender and Nomad, and also hooked it up to directly import to both Unreal and Unity and it's been quite good. For the most part, I will require to clean up the meshes, but that's excellent because I'm not looking to replace my entire workflow but to enhance my toolset, as I still do love modelling and being in creative control. It's an amazing tool for a 'base' if anything.

Performance-wise, so far, so good. There are some scenarios where it struggles, for example I tried creating a simple shield with multiple references and it was insisting on including an entire arm on which it's attached to, across 8~ attempts or so. At that point I gave up and just modelled it myself and then uploaded the model and let Meshy work with what I had made for it instead. At other times, it's the complete opposite. From the shield I made I used it as a reference to create a sword and the sword was absolutely fantastic. It honestly surprised me. It's all about finding a workflow and getting to know the tools because it's super fun once you get it going!

As for pricing; I'm pleased. I have admittedly tried some other services, albeit not extensively, but Meshy has been the only service so far that's kept me coming back. Full transparency; it's mostly been because they have been consistently offering special deals to continue subscribing, unlike other services I've tried. I'm a student so money is a concern, and in this economy? Yeah, it definitely matters, and Meshy simply provided what I need for the price I can afford so... I'm happy with that.

Support? Well, I actually have been in contact with them. I had an issue at one time where instead of adding tokens to my account, it was removed from it. Getting in touch with the support however was extremely easy and quick and they were super helpful and we got the issue resolved without any problems. I hope they discovered a bug they could sort out from my unfortunate mishap, but it's fine. It's all fixed now! Nothing but good on that end.

As for the capabilities of the AI model(s) themselves. It's hard to really say. It's not exactly a chatbot so there's not much conversation to be had, but in terms of output, I think for the most part it's good. As I said earlier, it's a bit hit or miss sometimes, but it also very much depends on what you are doing. Animation is something I've not dabbled much in, so I can't speak for its quality. The texturing is kind of weird, sometimes the UV maps are excellent, sometimes a jumbled mess (but the result looks good on the model itself), although I do wish the UV maps could be properly handled at times since it does cause some shader-issues in game engines if not done correctly.

Overall; happy with Meshy, and excited for its future. I think they're excellent, and there's always new things being developed which is nice. Review collected by and hosted on G2.com.

Oh, I covered what I disliked earlier a tad, but to be more specific; I'd say it's for the most part the UV-mapping quality, as the UV islands can be extremely jumbled and difficult to work with.

I do also wish there was a way to specifically target a 'part' of the output mesh. For example in my previous example, I generated a shield and it kept having an arm attached to it for some reason despite my references images not having an arm. It'd be nice if I could prompt it to specifically target that part of the mesh and refine it as such, to enable an iterative workflow rather than a "One-click-process".

As nice as "One-click-processes" can be, I do prefer being able to choose when I want to be more precise and specific.

And while I don't have a lot of experience with the animation tools, it'd be nice with some means to manually tweak bones or parts as well. Basically, it mostly boils down to having precise control to adjust things!

And, this isn't so much a 'dislike'; I want to see more improvements to the low-poly mesh generator! I know it's in beta so it's coming, but think of this more as support for it. :) Review collected by and hosted on G2.com.

I’ve been using Meshy AI for a couple of weeks now and I’m really happy with the service. The image creations are high quality and work excellently with my Bambu A1 printer.

The platform is very user-friendly, it was quick and easy to get the hang of the interface and workflows. Whether you’re generating new concepts or refining ideas, it delivers reliable outputs that translate well into printable models.

Their customer support and overall experience have been smooth. If you’re into 3D design or printing, I highly recommend giving Meshy a try. Looking forward to creating more with it! Review collected by and hosted on G2.com.

I would like to be able to give text prompts along with my uploaded image. Review collected by and hosted on G2.com.

The API is by far the best feature. One of the best use cases we have is an interactive python script that iterates through a directory of concept art and uses Meshy's API to automate 2D->3D generation. This allows for better organization, and creates an efficient workflow for manual review to better see the vision before decided which concepts to move forward with. Review collected by and hosted on G2.com.

What we dislike most about Meshy is the web interface itself, particularly the file/asset organization. and this is not unique to Meshy. All of the major players in this field have quirks with their webpages. Review collected by and hosted on G2.com.

The speed at which usable 3D models are generated from simple text or image prompts is impressive. It's especially useful that the models can be exported directly for 3D printing, and the topology is usually clean enough to work with without much additional cleanup. The different style options (e.g. low-poly vs. realistic) are also helpful for quickly creating prototypes for different purposes. Review collected by and hosted on G2.com.

With more complex or highly detailed prompts, accuracy still has its limits – fine details sometimes get lost or misinterpreted. Texture quality also isn't always consistent, so manual touch-up in a separate tool is often needed. Generation time for more complex models could also be shorter. Review collected by and hosted on G2.com.

I primarily use Meshy to create 3D printable figures, and what I like best is the sheer speed of the generation process. It takes creative ideas and turns them into solid, 3D-printable models in mere minutes, saving me hours of manual sculpting time. It is an absolutely fantastic tool for rapid prototyping and bringing custom figures to life effortlessly. The user interface is very intuitive, making the whole workflow incredibly smooth. Review collected by and hosted on G2.com.

Sometimes, the generated models require a bit too many revisions to get the small details exactly right. Depending on the complexity of the prompt, I might have to generate a few variations before getting the perfect result. Review collected by and hosted on G2.com.

The new Topology mesh feature in particular is really good for optimized meshes that are almost game-ready on their own. Many assets are absolutely game ready, but characters or clothing need a little extra work outside of Meshy. Review collected by and hosted on G2.com.

I would like to be able to actually set the smoothing for the normals, so that exporting a topology mesh can already be smoothed. Right now I can select flat, smooth, or auto (auto is best) for rendering in the preview, but it does not apply anything to the model, so exports come out flat and I have to smooth the meshes manually in Blender after exporting. Review collected by and hosted on G2.com.