Introducing 'Dematchi,' the official mascot of our Kyoto office, fondly known as Demachi Gadget! Thanks to the power of GPT-4o, Dematchi can now chatter away in fluent Kansai dialect! Enjoy the charming combination of our LOVOT robot with AI in a traditional Japanese house infused with technology. And please feel free to come see him if you're around. 🎙️

Our Tech Director, Kenta, has been deeply researching the capabilities of Three.js and WebGPU. His exploration of these technologies has uncovered fascinating insights, leading to the creation of a short article (in Japanese) that explains the changes and developments in this field. https://zenn.dev/wtshm/articles/2d55e77f48de55

Utilizing GLSL to experiment with raymarching techniques, Kenta has created a visually impressive demo based on his research. The result? A stunning demonstration that showcases the potential of raymarching in 3D graphics.

Check out "State," a social media platform that completely eliminates hierarchies.

"State" is a personal project planned and developed entirely by Qanta Shimizu, Co-Founder of BASSDRUM. He has been continuously updating it, and now you can request an invitation code on the website. Feel free to join!

3D printing technology has advanced significantly! Take a look at what our Technical Director, Takuma, has created. He used the Bambu Lab A1 mini to print our mascot character "Dramee." By pairing it with AMS lite, which allows for multi-color printing, he was able to craft a palm-sized and adorable figure right at home, all within 24 hours. The colors and details are accurately reproduced, resulting in a satisfying level of completion.

Utilizing MetaGPT to Enable Collaboration Among AI Agents with Different Roles, we requested to create BASSDRUM's website to assess its capabilities! Although the results were somewhat mixed with issues like missing Package.json, isn't it adorable to see multiple agents working simultaneously on the screen?

Using text-to-speech AI technology allows us to generate various voices with a unique speaking style that fits our needs. Since we're aiming to transform Dramee into an AI agent capable of supporting Tech Directors, we've been exploring various text-to-speech AI technologies to see if we can create vocal cords for Dramee.

Which voice do you think suits Dramee best? Also, if you know of any other services that might be suitable for generating his voice, please let us know!