XAgent

The goal of the agent is to automatically solve various tasks. The agent is developed for a wide range of tasks. It functions in a secure environment, using Docker containers to perform actions and manage tools such as file editors, web browsers, and interactive Python environments. XAgent supports collaboration with humans and is designed for extensibility, allowing users to add new tools and agents to improve its capabilities.

The agent can perform complex tasks without human involvement. It can work in a collaborative mode, following user instructions when solving complex problems, as well as requesting help when encountering difficulties.

To run the agent, Docker and Docker Compose must be installed. To start, dependencies (Python >= 3.10) must be installed, and XAgent must be configured in the assets/config.yml file, providing an OpenAI key. It is recommended to use models gpt-4-32k, but gpt-4 and gpt-3.5-turbo-16k are also acceptable as backup models.

Access to the agent is provided through a graphical user interface (GUI) or the command line. For interacting with the agent via GUI, a web server running on port 5173 is used. The standard login credentials for entry are guest and xagent.

The agent can use the following tools: a file editor for text manipulation, an interactive Python notebook for validation and plotting graphs, a web browser for information searching, a command line shell for executing system commands, and the Rapid API tool for calling external APIs.

The agent is an open project with the Apache-2.0 license and is available for free.