Inspiration
As someone who regularly models in CAD and 3D prints, I know the frustration of having a print fail because I didn't catch an issue beforehand, or because a part was too thin. That's why I built PrintGuide AI: to evaluate designs before printing and help users catch potential problems before wasting time and 3d printing filament.
What it does
PrintGuide AI reads an STL file's geometric data and takes a frame of the model, then analyzes it using Groq. The AI provides an overview of what to watch out for, what's done well, recommended print settings, or answers to something specific the user asks.
How we built it
The frontend of this project was built using HTML and CSS. For the 3D model viewer, Three.js was implemented so users could view their models directly in the browser with orbit controls. Node.js was used for server management, AI integration, and managing the AI chat. For our AI, we used Groq's Qwen 3.6-27B model as our main model and openai/gpt-oss-20b as a secondary model.
Challenges we ran into
One of the biggest challenges was prompting. I had to tweak the initial prompt many times because the AI wouldn't always provide enough useful detail or would sometimes sound too dry and robotic. Finding the right balance between technical accuracy and a helpful, natural response took a lot of iteration.
Accomplishments that we're proud of
One accomplishment was successfully integrating AI into the web server. This was my first time combining frontend development with AI, so getting the entire system working together was a major accomplishment for me.
What we learned
I learned how AI can be integrated into a frontend application and how prompting can give an AI a specific "role" and purpose for the user. I also learned how Three.js can be used to easily view and interact with 3D models directly in the browser.
What's next for PrintGuide AI
Next, I want to improve the AI chat area by adding chat history and making conversations more useful over multiple questions. One major feature I want to add is the ability for the AI to actually highlight critical points on the 3D model that are prone to stress, weak geometry, or printing problems.
Log in or sign up for Devpost to join the conversation.