Create multimodal AI agents

Step 1: Fill in the necessary info regarding your agent: its name, personality, skills, functions, and AI model

In this step, you can choose to build up the agent character based on 4 different sources

• a system prompt, that outlines its personality, purpose, and how it can help you or other users

• a link ( or many links) as knowledge base

• A document ( pdf file, note, excel sheet, and other formats) that act as the main knowledge base and the entire agent is based on that knowledge.

• an X profile, if you wish to create an agent out of a X profile personality, ( aka a x CLONE) , just past the username and the system will scrape the entire tweets of the person/account, and will train the agent on those tweets as its knowledge base

Step 2: Add any specific knowledge base( optional) and choose if you want it as a private agent, or public

Step 3: Choose the language, Choose a voice, or clone one ( even yours)

You can select from a significant number of languages, and you can choose a voice from a list of about 1k differnt voices, or clone your own. Voice is an important element in the multimodality element, as we believe software needs to be used in the most natual language, which is voice. once done, click on finish your companion

Step 4: Your agent is alive

Your agent is up and running, alive, and ready to produce, ( or just chit-chat). and by default, you can access all the different capabilities, with image generation, website generation, and web browsing If you want to create a tradable agent, check this page

Create tradable agents

Tokenization of AI agents and AI models