If you’ve ever wanted to have a personal chatbot that responds just like a real human, now you can! SmartestChild is an AI chatbot that has learned from thousands of internet conversations to become more like you, a user. This chatbot goes through a large language model that runs directly in your browser using WebGPU, and don’t worry—no data is sent to the server. Let’s talk emojis, abbreviations, and free expression!
SmartestChild functions by running off a Large Language Model (LLM) named Vicuna-13B. When you first land on the page, if it’s your first time, the model gets cached in your browser to boost performance. With WebGPU enabled, it brings this powerful language model to life—quickly responding to your queries and replicating natural human exchanges.
Powered by Apache TVM and MLC Relax Runtime, this ensures that all the hard work of understanding language and making it internet-friendly happens right in your browser without any server interaction.
Tweeting about it? SmartestChild has a Twitter account you can follow for the latest updates. And if you’re technically inclined, you can also view the project’s code on GitHub to see the magic for yourself.
In summary, SmartestChild is the internet-savvy, empathetic chatbot you’ve always wanted. So go ahead, open the link and start chatting in internet language, abbreviations, and offbeat emoticons. SmartestChild is here to be your personal internet pal!