Chatty AI is an AI assistant driven by the GPT model, while also combining open-source NLP models (such as GLM-130B, Vicuna, LLaMA-Alpaca, etc.) to provide comprehensive tiered AI intelligence. It is positioned as a RAG AIGC service with metaverse elements, providing customized AIGC services.
In terms of main capabilities, Chatty AI supports conversations with thousands of AI virtual characters. The characters cover professional roles such as IELTS teachers, project managers, and diplomats, and also include historical, cultural, and game characters such as Li Bai, Trump, and Raiden Shogun. They can be used to answer questions, complete work tasks, or learn about history, culture, game backgrounds, and anecdotes. In writing and daily creation, it can help complete tasks such as summarization and expansion, weekly report assistance, universal naming, replying to the boss, Xiaohongshu/Douyin/Weibo/WeChat Moments copywriting, typo correction, news commentary drafting, multilingual paper polishing, paper plagiarism reduction, and article expansion. The learning assistant function provides support such as assisted reading, understanding, and analysis of academic papers, cultivation of learning skills, formulation and completion of study plans, after-class discussions, and homework tutoring. In development assistance, it can generate code such as SQL, front-end components, Restful APIs, and operations scripts, and supports code analysis, code comments, and test case generation. Intermediary consulting services cover visa consultation, study abroad consultation, legal consultation, terminology consultation, business consultation, bidding processes, license consultation, company consultation, etc. The AI will provide customized analysis based on the information provided by the user, and plan the application plan in combination with standard general processes. It can also provide specific cases and simulated scenarios.
In terms of usage, Chatty AI provides document-style conversations. Users can edit input content in Markdown format, supporting multi-line text, headings, paragraphs, links, tables, mathematical formulas, code, etc. The edited content can be previewed in real time; the send key can be customized to prevent accidental touches when inserting line breaks. Users have control over their data. The platform provides session cloud storage services, allowing work progress to be shared across multiple devices, and works with a local storage mechanism. The product supports access on multiple devices such as mobile phones, computers, and tablets, supports free switching between Chinese and English, and all AI models support multilingual conversations, with language understanding and translation functions. Customer service support is provided through channels such as WeChat, email, and feedback.
Applicable scenarios include creative problems in daily work and life, academic learning, code development, document application consultation, and virtual character chatting, etc.
In terms of limitations, the materials mention that different commercialization packages correspond to different AI levels, but do not provide specific package names, prices, or feature differences; conversation migration and local storage mention twospecifications, 1000 items and 4000 items, but do not explain the corresponding packages or how to obtain them; the materials state that the code generation accuracy rate is guaranteed to be above 90%, but do not provide testing conditions and scope of application. Other functions, fees, platforms, or effect information not explicitly stated in the materials are not yet provided.
