I don't see many new models since I last checked the site a few months back. Traction?<p>The problem is probably that most good 3d modelers do not need your service. So why would they register? And even pay for it? What'd their incentive be?
One would be that they can earn extra by selling models on your site. But there is Turbosquid & co already doing this since over a decade.<p>There's a chicken-egg issue here: how do you get people to go to the website with their CC ready to buy stuff from the few artists who will have registered by then? You probably don't.<p>The short version is that very likely, displaying 3D models in a web browser, even if interactively, and even if everything is looking really sexy, as it does on your site, is not a viable business model per se. :)<p>Most 3d models are used in desktop apps, not on the web (even if they end up on the web at some stage, much later).
And what Turbosquid & co haven't got is a direct channel for the stuff on their website into the content creation apps.<p>So that is where you can get leverage. Write plugins for Maya, Max, XSI, Modo, Houdini, etc. that allow browsing and loading models live into the app. As this is quite a task, the order of developing the plugins should strictly be by market share of these apps.<p>Then you need lots of free high quality models as an incentive for people to install these plugins.
I would just skim all stuff available online, categorize it and put the best into your catalog.<p>That will get you traction on the customer site ("customer" here refers to people willing to buy models from the artists who register on your site) and then, as a result of that, you will really get artists to consider opening a portfolio on your site (and maybe, even pay for that). You could then also think about a business model that is purely based on taking a % of the sales of an artist, instead of a fixed, monthly price.<p>The plugin would also allow inspecting models directly inside the app. If the model is paid content, the plugin would only load 50% of the geometry until the model is paid for. This is still plenty to assess quality.<p>Basically, get people to make the plugins part of their everyday workflow, then you will get them to consider taking their CC cards out if they need to load an asset that is paid content.<p>And here is a really cool feature that becomes possible through the plugins: allow the artist to push updates of the models to the client who bought the right to use it, right into the desktop app. Basically if I buy a model and I need a slightly modified version and the artist agrees to do it, they can do it live.
Then you can even offer billing by the hour and take crop. etc.<p>Plus you give people in need of models a direct line to artists. Most of the time when you quickly need a model, it is something specific. Even if you have the money, you need to find someone who does it at the quality you require and matching the style/art direction you got from your own client.
How nice would it be if one can put a placeholder bounding box in their 3d scene, attach a description of what they need in there to it (plus maybe some Dropbox links with art direction/images) put a price out they are willing to pay and go to sleep. When they open the app the next day they may already have some rough models from some artists who are bidding on it to cycle through, right in their 3d scene.<p>Check out what designcrowd and similar websites offer for 2D. There's nothing really like this for 3D. And the plugins will ensure no one can copy your business model easily (or quickly).<p>On the site (and inside the plugins) you could offer a variety of services that could be up-sold on top of the base price of the model (and the net go straight into your pocket). Some ideas:<p>- Automatically check models for manifold topology, mark models as non-manifold if they are or allow fixing on the fly.<p>- Conversion.<p>-- An artist may upload a high detail NURBs model of a car but a client needs a meshed polygonal version. You could do the meshing on-the-fly, even supporting LOD. AYAM is an OSS NURBs modeler that uses OpenGl, so it needs to mesh the NURBs, pinch code from there. Supporting OpenNurbs as a format should be fine for starters.<p>-- A polygon mesh with smoothing groups can be converted into a subdivision surface with edges and vertices auto-tagged as creases and corners, based on continuity. 3Delight for Maya uses this approach. OpenMesh, a HE data structure lib has feature detection functionality making this stuff a few dozen lines to implement.<p>- 3D printing. This requires meshing. Both Aqsis and Pixie are OSS micropolygon renderers that can be made to dump micropolygon grids to watertight STL files ready to print. Get a deal going with Shapeways, i.materialize etc. or all of them so people can order prints straight through your website.