Advertflair, a 3D studio, asked all their questions about Convrse.ai so you don’t have to

Vikrant, Founder, Convrse.ai recently spoke to Hari and Praful from Advertflair and they had tons of questions. Read on to know about tech behind Convrse.ai, how it works and why it is a game-changing platform for 3D creators.

Advertflair, a 3D studio, asked all their questions about Convrse.ai so you don’t have to

There is a special place in heaven for 3D studios who work with clients on customer facing interactions. Delivering hyper realistic experiences that clients expect today with lean teams, leaner budgets and tech that needs serious advancements (still touted as the best tech available) is a miracle they perform everyday. But, this is a tough act to scale without building serious intelligence and automation in workflows, including in optimization.

Studios know it. And they are generously helping ecosystem builders like us to understand workflow challenges, and build smart optimization solutions for each step of the creator journey. In fact, much of intuitive intelligence in Convrse.ai around scalability and efficiency has come from 100s of conversations with studio creators who are most hard pressed for time and budgets.

My conversation With Hari Senthil Gurusamy and Praful P Akella from Advertflair, a boutique 3D modeling studio

I recently spoke to Hari and Praful from Advertflair about Convrse.ai and they had tons of questions. As I answered these questions, I realized our conversation could actually help many other studios wrap their heads around Convrse.ai and what we mean when we say that it is game changing. So, here goes:

The Challenge Faced by Advert Flair

Advertflair struggles with compressing the extensive, hyperrealistic inputs, often reaching sizes as large as 200MB, particularly GLBs housing intricate 3D models. These models not only serve VR/AR purposes but also play a crucial role in enhancing user experience and interaction through animations.

Another crucial challenge is to solve workflow integration with new clients who resist investing in their inbuilt systems. To address this, they aim to offer a SaaS solution for clients to seamlessly integrate into their existing backend processes

We went through Convrse.ai together and Hari and Praful had the following questions that I answered. It was a great exchange - always is when I connect with fellow 3D enthusiasts!

Top questions 3D studios have about Convrse.ai

💡
Hari: Do you offer APIs? And do you provide white-labeled solutions that we can push at our client end?


Convrse.ai Capability: Yes, we do provide both an API and white label solutions for our clients. Our main focus is on building a bridge to convert legacy CAD and traditional 3D data into suitable file formats such as GLBs, FBX, and OBJs. We optimize these files by removing hidden objects, emphasizing surface layers, and enhancing photorealistic visual representation while preserving the integrity of the original mesh. Additionally, we offer APIs for seamless integration, and we are currently piloting this solution with a leading ecommerce platform.

💡
Praful: How does the system decide the number of vertices in 3D models and prioritize them based on importance? Is there a specific algorithm involved, or is it based on an AI-powered machine learning model trained in a particular way?

Convrse.ai Capability: We've trained our system using a standard graph neural network algorithm around mesh optimization in 2022. Subsequently, we trained it on our own data, including the heavy models we received from architects in formats such as CAD, SketchUp, and Revit, as well as the optimized models we generated. The training process was based on our manual approaches in the past, aiming to automate the optimization algorithm. As a result, we've developed a comprehensive automation engine that replicates human actions automatically.

💡
Hari: How do we integrate into existing workflows? Our workflows are process-driven at scale, often multilayered, encompassing compliance, internal quality control, and external quality control. Is there an API solution available? If so, how does it function?

Convrse.ai Capability: It's quite simple. When a mesh is sent to us, you can automatically define parameters. For instance, if your platform has a limit of 50MB and 200K triangles, you can specify these limits for all your inputs. Therefore, when your customers upload data to be sent to Convrse, what you receive will always be 50MB or less and have fewer than 200K triangles. Moreover, you have the option to utilize your own user interface, seamlessly integrating our services into your system. Many people are using us as a standard import channel for their platforms. The process typically involves the customer uploading the file, defining texture, mesh size, and format on your platform, specifying the desired format for conversion. Once you've done this, the file is sent back to us, optimized, and returned to your VR editor.

💡
Praful: Is there maximum flexibility regarding those parameters?

Convrse.ai Capability: No, platforms are free to set different limits for maximum optimization. One platform might want to keep it at 100K, 500K, 10K - you decide that. The only restriction could be a user’s RAM when it's being consumed on the browser. On MAC, you can go up to 4MN polygons and it works well.

💡
Hari: We can request customization based on a preset? If we want the output to be a certain way, can you give us a quality control check?



Convrse.ai Capability: We are building that layer, where we are building an engine within Convrse.ai to identify the level of shape deformation. Beyond a point, it will stop deforming the meshes by identifying that the mesh has been played around to maximum permissible limit and it will stop optimization. We are building that QC layer.

💡
Hari: Can you provide training to our team members?

Convrse.ai Capability: Definitely! We will set up a training session and integrate it as a part of their workflows. They’ll love it because with a plugin we can create a bridge between their preferred modeling software (3DS max or Blender) and their preferred gaming engine -they might not even need to exit their usual platform of work

Praful: This is exciting! Integration as a plugin within Blender and 3ds Max would seamlessly incorporate our tool into the entire workflow, making it the go-to solution.  From the client's perspective, this optimized experience would justify any associated cost adjustments!

Impact of Convrse.ai

Over 3500 game developers that use Convrse.ai, about 500 come back every week
9217 total models and scenes optimized
25 work days saved on an average for our users
$1000 dollars saved on manual work on an average
2.3 BN polygons reduced for creator community globally

If you have any other questions about Convrse.ai as a studio, an indie creator or an agency - feel free to send them to me. We are working hard to make Convrse.ai work for the 3D community - your questions help us guide in the right direction.

🙋🏻‍♂️
Say hello to the author: Vikrant Singh, Founder & CEO, Convrse.ai

Vikrant is a seasoned entrepreneur, a 3D nerd and CEO at Convrse.ai. He has built and managed many small ventures, wearing multiple hats in business development, product development and sales & marketing. Amongst the accomplishments he holds close to his heart are creating over 200 immersive experiences for brands in the VR industry; winning the Best VR Startup award in 2021; and being the only Web 3.0 startup featured in the prestigious Your Story Tech30 list. At Convrse.ai, Vikrant leads efforts to revolutionize digital interactions for the new era of internet with the use of real-time 3D. Vikrant is a graduate from IIT-Guwahati.

World's smartest 3D model optimizer

Convrse.ai is a cloud-based 3D optimizer that supports 30+ file formats including CAD files and comes with a file viewer, heatmap and a unique selective optimization feature.

Try Convrse.ai