Overview

In TaskingAI, an Assistant is an AI agent whose functionalities can be customized according to user requirements. It operates by being bound to a chat completion language model available within a project. Users define the Assistant's tasks and roles through a system_prompt_template, which guides its behavior and response mechanisms. Memory integration is a key aspect, providing the Assistant with the capability to store chat contexts, essential for maintaining continuity in conversations.

For advanced functionalities, Assistants can be equipped to perform Retrieval Augmented Generation (RAG) by integrating with retrieval collections. This enables them to access and use information from a user's private database for answering queries or engaging in informed conversations. Additionally, Assistants can be integrated with various tools to gain the ability to interact with external resources and perform tasks like checking real-time flight information, weather updates, or sending specific communications.

In this section, we'll introduce the basic operations to manage Assistants, such as creating, updating, retrieving, and deleting Assistants.