Build AI chatbots with Coze! All experience levels welcome.

The live stream begins with an introduction to building AI chatbots using Cod AI's no-code or low-code platform. The host, Zara, opens the comment section and welcomes viewers to the stream. She explains that they will be creating a 'Chad bot' from scratch, showcasing Cod AI's features and answering questions from the audience. The stream is suitable for developers of all levels, and viewers are encouraged to create a free account on Cod AI's website to follow along. Zara and Joshua, a developer advocate for Cod AI, introduce themselves and their roles. They also discuss the global audience present in the chat, highlighting the platform's worldwide reach.

The hosts discuss the bot store, which contains community-created bots for various purposes, such as learning, productivity, and entertainment. They demonstrate how to interact with a GitHub expert bot, which uses plugins like GitHub and browser integration to provide information. They also show how viewers can customize and duplicate bots through public configuration. The conversation then shifts to the differences between Cod AI and other AI bot builders like GPT Builder, highlighting Cod AI's additional features such as knowledge bases, workflows, and multi-step processes.

The live stream continues with the creation of a 'Plan a Trip' bot. The host guides viewers on how to use persona prompts to design the bot's character and functionalities using natural language. They also show how to optimize prompts for better understanding by the AI and use markdown formatting to emphasize certain characteristics. The host explains the use of large language models like GPT 4 8K and how to select and customize them according to the bot's needs.

The hosts test the 'Plan a Trip' bot by asking it to recommend high-rated Ramen restaurants in Ginza, Tokyo. They discuss how the bot uses GPT's training data to answer questions and how to improve its responses with real-time or privately available information. The hosts introduce plugins, which are applications that provide additional functionalities like web searches, text-to-image generation, and integration with services like Trip Advisor and Yelp. They demonstrate adding these plugins to enhance the bot's capabilities.

The hosts delve deeper into the use of plugins, showing how they can be added to the bot to utilize external services without coding. They discuss the ease of adding plugins like Google web search and DALL-E for image generation. They also create a Trip Advisor plugin to access various functions from the Trip Advisor API, emphasizing the no-code aspect of Cod AI. The hosts explain how to customize bot responses using plugins and how to create personalized plugins using external APIs.

The hosts introduce card bindings, a feature that helps structure bot responses in a visually appealing way. They demonstrate how to use card bindings with the Yelp API to present restaurant recommendations in a card format, including title, description, and image. The hosts also show how to customize the card binding data to match the desired response format and how users can interact with the cards to access the recommended pages.

The hosts discuss how to make the travel bot more personalized using variables. They guide viewers on creating variables for trip preferences, which can store user-specific information like interests in hiking, historical sites, or food. The hosts explain how the bot can use these variables to provide customized responses based on user preferences, enhancing the user experience.

The hosts introduce knowledge bases, which allow the bot to reference content from various sources like files, websites, and APIs to provide more contextually accurate responses. They demonstrate how to create a knowledge base and add different types of data sources to it. The hosts also explain how to use the knowledge base to create a retrieval-augmented generation (RAG) pipeline without coding.

The hosts explore multi-agent mode, which enables the creation of a team of bots, each specializing in different skills to work together to complete tasks. They provide an example of using multi-agent mode to create a personal assistant bot that manages email, calendar, and Google Sheets. The hosts discuss how this mode can be used to create more complex and efficient bots by combining different skills and functionalities.

The hosts guide viewers on how to publish their bots to make them shareable with others. They discuss the process of publishing to the Cod AI bot store and other popular chat applications like Discord and Telegram. The hosts also mention the benefits of publishing to the bot store, including increased exposure and the ability to feature the bot on the store's recommended section.

The hosts provide information on additional resources available for learning more about Cod AI, including comprehensive documentation, the Cod assistant bot in the Discord server, and a YouTube channel with tutorials and case studies. They encourage viewers to join their Discord server for further questions and feedback. The live stream concludes with thanks to the participants and an invitation to explore and build on the Cod AI platform.