Server

AgentMode is a full - featured Model Context Protocol (MCP) server that provides an integrated data integration solution for AI development by connecting to multiple databases, data warehouses, cloud services, etc.

Developer tools Database #Data integration #AI development #Cloud services .Python

rating : 2.5 points

downloads : 7.4K

update time : 2025-07-31

Open Site

What is AgentMode?

AgentMode is an intelligent development assistant server based on the Model Context Protocol (MCP). It allows your AI programming assistant to directly access various resources such as databases, data warehouses, and cloud services, significantly improving development efficiency.

How to use AgentMode?

Install it quickly through the VS Code extension or manually configure it using the Python package. After starting, configure the data connection through the browser interface, and the AI assistant can directly query the data source.

Applicable scenarios

Development scenarios that require an AI assistant to access multiple data sources, such as data analysis, cross - platform data query, cloud service integration, and automated data processing.

Main features

Multi - data source connection

Supports databases such as MySQL and PostgreSQL, and data warehouses such as Snowflake and BigQuery

Cloud service integration

Will soon support mainstream cloud platforms such as AWS, Azure, and Google Cloud

Automatic dependency management

Automatically installs the required dependency libraries when configuring the connection

Advantages

One - stop connection to multiple data sources

Local storage of credentials for security

Automatic handling of dependency relationships

Visual configuration interface

Limitations

Some cloud service functions are still under development

Requires basic configuration knowledge

How to use

Install the extension

Install the AgentMode extension in VS Code or install it through the Python package

Start the server

Start the MCP server through the command panel

Configure the connection

Access localhost:13000/setup to configure the data source connection

Usage cases

Database query automation

The AI assistant directly queries the production database to obtain the latest data

Cross - platform data comparison

Simultaneously query MySQL and Snowflake for data consistency check

Frequently Asked Questions

Where are the credentials stored? Is it secure?

Which databases are supported?

Related resources

Official documentation

Detailed installation and usage guide

GitHub repository

Source code and issue tracking

Discord community

Get real - time help and communicate

🚀 AgentMode ✨

AgentMode is an all - in - one Model Context Protocol (MCP) server. It connects your coding AI to dozens of databases, data warehouses, data pipelines, cloud services, and more. This extension streamlines your development workflow by providing seamless integration with various data and cloud platforms.

flow diagram!

🚀 Quick Start

💻 Installation

Quick start with VS Code

Install our VS Code extension.
Click the 'Install' button next to the AgentMode extension.
Start the MCP server via the Command Palette (Ctrl+Shift+P or Cmd+Shift+P on macOS), then type in 'MCP', then select 'MCP: List Servers' and click on AgentMode.
Click on the 'AgentMode' text in the bottom - right status bar corner of VS Code to open a browser tab, sign in, and then setup your connections.

Without VS Code (Python package)

Open the terminal and install uv with curl -LsSf https://astral.sh/uv/install.sh | sh
uv init
uv add agentmode
uv run agentmode

MCP configuration for VS Code

If you installed our VS Code extension, it will automatically create or update your settings.json file for you. If you installed AgentMode manually, please create a .vscode/settings.json file in your workspace, and replace ‘YOUR_INSTALLATION_FOLDER’ with the path to your uv environment below:

{
    "mcp": {
        "servers": {
            "agentmode": {
                "command": "cd 'YOUR_INSTALLATION_FOLDER' && uv run agentmode",
                "env": {}
            }
        }
    }
}

MCP configuration for Cursor

Please create a \~/.cursor/mcp.json file in your home directory. This makes MCP servers available in all your Cursor workspaces. Please replace 'YOUR_INSTALLATION_FOLDER' below with the folder you setup your uv environment in:

{
    "mcpServers": {
        "inputs": [],
        "servers": {
            "agentmode": {
                "command": "cd 'YOUR_INSTALLATION_FOLDER' && uv run agentmode",
                "env": {}
            }
        }
    }
}

MCP configuration for Windsurf

Open the file ~/.codeium/windsurf/mcp_config.json Add the code below to the JSON file. Press the refresh button in Windsurf. Please replace 'YOUR_INSTALLATION_FOLDER' below with the folder you setup your uv environment in:

{
    "mcpServers": {
        "inputs": [],
        "servers": {
            "agentmode": {
                "command": "cd 'YOUR_INSTALLATION_FOLDER' && uv run agentmode",
                "env": {}
            }
        }
    }
}

🌐 MCP (Model Context Protocol)

AgentMode leverages the Model Context Protocol (MCP) to enable your coding AI to:

Access and query databases and data warehouses.
Interact with data pipelines for real - time or batch processing.
Use a web browser.
See logs from your production services.
Connect to cloud services for storage, computation, and more.

🔌 Connections

connections setup!

AgentMode supports a wide range of connections, including:

Databases: MySQL, PostgreSQL, etc.
Data Warehouses: Snowflake, BigQuery, Redshift, etc.
Data Pipelines: Airflow, Prefect, etc.
Cloud Services: AWS, Azure, Google Cloud, etc. (coming soon!)

To configure connections, follow these steps:

Start the MCP server and go to http://localhost:13000/setup
Click on the icon of the connection you'd like to setup.
Fill out the connection details and credentials (all credentials are stored locally on your machine).
Any required dependencies will be installed on - the - fly.