Artificial intelligence

MyShell, a Multi-Modal Chatbot Platform, Revolutionizes the Interactive Frontier for Multimedia Experiences

World-class alumni and AI developers, Ethan, Rick, and Charles, have partnered to launch MyShell, a codeless, user-friendly chatbot platform with multi-modal functions. According to the co-founders and their team, who hail from prestigious universities like Oxford, MIT, Princeton, and Tsinghua, MyShell provides users with unparalleled control over their brand experiences.

While traditional chatbots rely on a combination of large language models (LLM) and pre-defined prompts to generate responses, they fall short when handling more complex tasks such as realistic text-to-speech and image generation. Another observable shortcoming is their lack of humanization and warmth during interactions. Their responses are robotic and monotonous, and they don’t cater to specific industries or professions.

MyShell’s public release is a game-changer since it breaks free from these limitations by utilizing human-like voice, fast voice clone, and knowledge base generation and translation technologies. These features enable realistic speech communication between users and the robot, which is useful for language learning, programming learning, storytelling, and other environments. They can train their robots using short voice clips and clone their characters for a more personalized experience.

The brains behind this feat have a history of developing complex, data-driven solutions for the consumer and enterprise industries. Ethan, an Oxford alumnus and serial entrepreneur, founded his first company during college, which was later acquired by KE Holding and became the most widely adopted VR application. He is also the co-founder of ATP.CLUB, an AI-driven fitness gaming platform and the Algorithm Leader in the robotics division of a renowned AI unicorn company.

Rick is notable for his 10 years of AI product experience. He is the founder of a start-up that has the most commonly used AR SDK among Chinese companies. Like Ethan, he is also the co-founder of ATP.CLUB and the leader of the robotics division at a renowned AI unicorn company. Charles is an MIT PhD grad with a background in robotics, computer vision, and reinforcement learning. During BSc study at Tsinghua University, he published over 8 top-tier conference paper.

The joint efforts of these co-founders to bring their vision to reality with the launch of their flagship product in March have been nothing short of rewarding. MyShell has gained over 100,000 users through community-driven growth and with zero budget in just two months. The platform boasts a strong community of over 10,000 members who contributed to its early development and organic spread and adoption, of which 13 are now Shell DAO major contributors, three of whom participated in seed investments.

MyShell’s steady progress is noteworthy, seeing that It started with the project initiation and creation of a simple chatbot demo on Telegram in March. Three weeks later, it piqued the interest of over 30,000 users, mainly comprising language practice and university student communities, many of whom were highly engaged in product development and marketing. During this time, series of testing revealed that the TTS algorithm achieved realistic results, reducing costs to 3% of any commercial solution.

To put this into perspective, Microsoft TTS solution requires 30 to 180-minute samples to generate 100 characters within 0.2 to 0.4 seconds. It costs users $24.04 to generate 1 million characters. Meanwhile, MyShell TTS only requires 3-minute minimal cloning of the same number of characters in 0.2 seconds. Plus users can generate 1 million characters for $0.92. Even better, they sound natural.

In April, the developers launched MyShell’s web version, the Vitalik bot designed with the Knowledge Base function and well-versed in the latest Ethereum development, and Leidian Shogen – a Genshin Impact character based on the Japanese TTS model. The platform’s user base also reached its 100,000 milestone.

May saw the launch of the Creator Workshop, which supports the development of custom chatbots with cutting-edge modules such as multi-language TTS and Knowledge Base, beyond Chat GPT capabilities. As of the 4th, the community had created over 100 chatbots with various functionalities. One notable launch was the popular playable character, March 7th, in Honkai: Star Rail, which is poised to provide game strategy Q&A.

In addition, the development team didn’t hold back in refining the Knowledge Base feature, after which it launched the Chinese TTS model. So far, the workshop and chatbot creator tools offer multi-modal integration, which allows creators to choose different technologies to implement and empower their ideas. The main categories of the user-generated content (UGC) robots on the platform are education, entertainment (from horoscopes to text RPG) and tools (including translation and proofreading, gaming strategies, and legal assistants).

Some remarkable bots built by the creators include: 

  • YUKI – IELTS Teacher Ben, an AI tutor that helps students prepare and take the IELTS test without breaking the bank. The bot provides personalized training that embeds language organization skills, preparation strategies, and feedback, among others. Its voice emulates the famous detective: Sherlock Holmes. 
  • Yellx, an RPG adventure game that helps users learn English during game play by generating corresponding game texts based on their specified English level, from CET-4 to TOEFL.
  • Yankairui, a cat-raising bot that serves as a daily companion and provides an in-depth exploration of the world of cats.
  • Siyuan – Leo & Allen, a dual personality bot that provides two different approaches to a topic.
  • Kaiserwetter, an MTBI simulation that generates any character of choice based on their personalities, including their age, identity, hobbies, and recent activities.

MyShell is positioned to incorporate more high-end features in the nearest future, including one-minute audio cloning, multi-modal AI and LLM, emotion-controlled TTS, and endless memory.


To Top

Pin It on Pinterest

Share This