Skip to main content

Integrate with External Tools

This page explains how to integrate Agentiqs with external tools.
You can integrate with a variety of tools, including Microsoft Office and GitHub, as well as other tools.

As representative examples, this page describes the following integration methods:


Integrate with Applications and Web Services

This section explains how to integrate Agentiqs with GitHub.

Procedure
  1. Click [Tools] > [Create New].
    • The [Tool Gallery] dialog appears.
  2. Select [GitHub] from the list.
    • The GitHub settings screen appears.
  3. Enter a Personal Access Token in the [GitHub API Key] field.
    • The credentials are saved, and you can connect to GitHub.
  4. Add the created tool to an AI agent.
    • In chats that use that agent, you can view and update GitHub data.
Available tools
  • For the tools available in Agentiqs, see Tool List.

Configure MCP Servers

Configure per Workspace

MCP servers used only in a specific workspace are created as tools.
MCP servers themselves are not provided by Agentiqs, so users must create and manage them.

Prerequisites
  • You have created and are running an MCP server that supports the MCP specification.
  • You know the information required to connect to the MCP server, such as the host name/port, connection method, and any required credentials.
  • The MCP server is reachable from Agentiqs over the network.
Procedure
  1. Click [Tools] > [Create New].
    • The [Tool Gallery] dialog appears.
  2. Select [MCP Tool] from the list.
    • The MCP server settings screen appears.
  3. Refer to the configuration example below and enter [MCP Server Settings].
  4. Click [Save] to save the settings.
Notes
  • Create and operate MCP servers at your own responsibility.
  • Agentiqs does not guarantee the operation or availability of MCP servers.

Configuration Example

In Agentiqs, [MCP Server Settings] should contain not the entire MCP server configuration file, but only the configuration for a single server you want to use under mcpServers.

Example of the full configuration file:

{
"mcpServers": {
"exampleServer": {
"command": "npx",
"args": [
"-y",
"example-mcp-server"
]
},
"anotherServer": {
"url": "http://localhost:9000/sse"
}
}
}

Example of what to enter:

When using exampleServer (mcpServers.exampleServer):

{
"command": "npx",
"args": [
"-y",
"example-mcp-server"
]
}

When using anotherServer (mcpServers.anotherServer):

{
"url": "http://localhost:9000/sse"
}

Notes on Configuration

  • On Windows, specify paths using one of the following methods:

    • Use /.
    • Write \\ as a double backslash.

Configure MCP Servers Globally in the App

Use the app-wide MCP settings when you want to use the same MCP server across multiple workspaces.

Once registered, it can be selected from all workspaces.

There is no need to create the same connection information for each workspace.

  • By writing JSON in mcpServers format, you can register multiple MCP servers at once.
  • If you already have MCP servers installed, you can import the settings.
Procedure
  1. Import from [Settings] - [MCP] using one of the following methods.
    • 1: When importing a new tool, enter JSON in mcpServers format.
    • 2: When importing an existing tool, click [Import from Other Tools] and select the automatically detected tool.
  2. Open the target AI agent and click [Tools] > [Select Tools].
  3. From [MCP (App-wide)], select the MCP server you want to use and turn on the necessary features.
Notes
  • If a tool with the same name exists in the workspace, the workspace tool takes precedence.
  • In shared workspaces, only the MCP server name is shared. Each user must configure the connection information.

このページは役に立ちましたか?