Easy Agents: Rapidly Building Intelligent Automated Agents Using Natural Language

Easy Agents is an open source framework designed to help users quickly build and deploy intelligent automation agents through simple natural language commands. It makes it easy for non-technical people to define complex workflows and reduce the hassle of writing code manually. Users only need to use natural language to describe the task , the framework can automatically generate the corresponding automation agent , applicable to security operations , customer support and IT management scenarios , etc. Easy Agents based on FastAPI and MCP (Model Context Protocol) , providing efficient tool integration and deployment methods . The official documentation provides four examples of how to automate complex tasks with simple commands. The project is maintained by Kyle Polley on GitHub and welcomes code and feedback from the community.

Function List

• Supports defining agent behavior in natural language without complex code.
• integrated (as in integrated circuit) MCP protocol to easily connect to external tools and APIs.
• Provides FastAPI endpoints and cron triggers to simplify agent deployment.
• Includes uv tools to manage project dependencies and ensure a consistent environment.
• Supports the creation of custom MCP servers to extend proxy functionality.
• Provides sample workflows to demonstrate scenarios such as security alert analysis, report generation, etc.
• Open source community support that allows users to submit issues and code improvements.

Using Help

Installation process

To start using Easy Agents, users need to clone the repository and configure the environment. Below are the detailed installation steps:

1. clone warehouse
   Run the following command in a terminal to clone the Easy Agents repository locally:

   git clone https://github.com/kpolley/easy-agents.git
   cd easy-agents
   

2. Installation of dependencies
   Easy Agents Usage uv tool to manage dependencies. Ensure that Python and uv, and then run:

   uv sync
   

   This will automatically install all dependent packages required by the project, such as FastAPI and related libraries.

3. Starting the Development Server
   After the installation is complete, start the FastAPI Development Server:

   uv run fastapi dev main.py
   

   Once the server has started, it can be accessed via a browser http://localhost:8000 View API endpoints.

4. Configuration environment
   If you need to connect to an external tool such as GitHub or Slack, you need to set the MCP server address in your project configuration file. Refer to mcps/github_extended.py Example, Configuring a Custom MCP Server.

Creating a Proxy

Creating a new agent is very simple and requires only the following steps:

1. Defining Proxy Behavior
   In the configuration file in the project directory, describe the agent's tasks in natural language. Example:

   你是一个负责调查安全警报的专家代理。任务步骤：
   1. 根据警报 ID 查询详细信息
   2. 深入分析警报内容
   3. 生成详细报告
   4. 如果警报有效，将报告发送到 Slack 的 #security-alerts 频道
   

   These instructions are parsed by the framework into executable workflows.

2. Specify the MCP server
   Specify in the configuration file the address of the MCP server that the agent needs to access. For example, to connect to the GitHub API, you can use the official GitHub MCP server or a custom server.
3. Deployment Agent
   Agents can be deployed via FastAPI endpoints or cron triggers. Run the following command to set up a timed task:

   uv run python scripts/run_agent.py
   

   or trigger agent execution via an API call.

Featured Function Operation

natural language programming

The core functionality of Easy Agents is to allow users to define complex tasks in natural language. For example, create an agent that generates weekly project reports:

每周日，代理会检查 Linear 的任务和项目，生成团队项目进展报告，列出本周完成的任务和负责人。

The framework automatically parses the commands, calls the relevant APIs (e.g. Linear API), and generates reports. The user does not need to write code, just describe the requirements.

Integration of external tools

Easy Agents makes it easy to integrate external tools through the MCP protocol. For example, connecting to Slack to send notifications:

1. Add the Slack MCP server address to the configuration file.
2. Include a step in the agent command to send a notification, such as "Send report to channel #team-updates".
3. Once the agent is deployed, reports are automatically sent to the specified Slack channel.

Customizing the MCP Server

Users can create custom MCP servers if the default tool does not meet their needs. Refer to mcps/github_extended.py, follow the steps below:

1. Create a new Python file that defines the MCP server logic.
2. Enables interaction with external APIs, such as fetching GitHub repository data.
3. Reference this server address in the proxy configuration.
4. Once deployed, the agent can invoke the customized tool.

Debugging and Extension

• adjust components during testing: The FastAPI server provides logs that record each step of agent execution. Users can check the execution of tasks through the logs.
• extensions: Users can extend the functionality by submitting a pull request or modifying the code. For example, adding new MCP servers to support other tools.

caveat

• Ensure that the network connection is stable and that the MCP server needs access to external APIs.
• probe uv Version compatibility, the latest version is recommended.
• Regularly update the repository for the latest features contributed by the community.

application scenario

1. security operation and maintenance (O&M)
   Easy Agents can be used to automate security alert processing. Agents can query alerts, analyze threats, generate reports, and send results to team communication tools, reducing manual time.
2. Customer Support
   Agents can automatically respond to common customer questions, organize support tickets, and generate summary reports to improve customer service efficiency.
3. IT Management
   For automated server monitoring, log analysis or resource allocation. Agents can perform tasks based on natural language commands, such as "check server status and send report every day".
4. project management
   Automatically generate weekly project progress reports to analyze task completion and notify the team, suitable for agile development teams.

QA

1. Is Easy Agents suitable for users with no programming experience?
   Yes, Easy Agents allows users to define tasks in natural language with no programming knowledge required. Non-technical people can get started quickly.
2. How can I extend the agent functionality?
   Users can extend the functionality by creating custom MCP servers or modifying the code. References mcps/github_extended.py Example.
3. What environments are required to deploy agents?
   Requires Python, uv tools, and an internet connection. Recommended to run on Linux or macOS.
4. What external tools are supported?
   Supports tools like GitHub, Slack, etc., and can be extended to any tool that supports APIs via the MCP protocol.