# Configure in a coding IDE
<a name="mcp-server-configure-ide"></a>

Coding IDEs that support MCP use a JSON configuration file to declare which servers to start. The server runs over stdio, so the IDE launches it as a subprocess when you open a project.

## Kiro
<a name="mcp-server-configure-kiro"></a>

Create or edit `.kiro/settings/mcp.json` in your project directory (workspace-level) or `~/.kiro/settings/mcp.json` (user-level, available in all projects). Because Kiro runs on your local machine, use `AWS_PROFILE` to pick up credentials from your existing `~/.aws/credentials` file:

**Amazon OpenSearch Service domain:**

```
{
  "mcpServers": {
    "opensearch": {
      "command": "uvx",
      "args": ["opensearch-mcp-server-py"],
      "env": {
        "OPENSEARCH_URL": "{{https://your-domain-endpoint.region.es.amazonaws.com}}",
        "AWS_REGION":     "{{us-east-1}}",
        "AWS_PROFILE":    "{{your-aws-profile}}"
      }
    }
  }
}
```

**OpenSearch Serverless collection:**

```
{
  "mcpServers": {
    "opensearch": {
      "command": "uvx",
      "args": ["opensearch-mcp-server-py"],
      "env": {
        "OPENSEARCH_URL":            "{{https://collection-id.region.aoss.amazonaws.com}}",
        "AWS_OPENSEARCH_SERVERLESS": "true",
        "AWS_REGION":                "{{us-east-1}}",
        "AWS_PROFILE":               "{{your-aws-profile}}"
      }
    }
  }
}
```

After saving, Kiro reconnects the server automatically. You can then ask questions like *"List the indexes in my cluster"* or *"What is the health of my cluster?"*

## Claude Code
<a name="mcp-server-configure-claude-code"></a>

Add the server to your project's `.mcp.json` file. As with Kiro, use `AWS_PROFILE` to authenticate with your local AWS credentials:

**Amazon OpenSearch Service domain:**

```
{
  "mcpServers": {
    "opensearch": {
      "command": "uvx",
      "args": ["opensearch-mcp-server-py"],
      "env": {
        "OPENSEARCH_URL": "{{https://your-domain-endpoint.region.es.amazonaws.com}}",
        "AWS_REGION":     "{{us-east-1}}",
        "AWS_PROFILE":    "{{your-aws-profile}}"
      }
    }
  }
}
```

**OpenSearch Serverless collection:**

```
{
  "mcpServers": {
    "opensearch": {
      "command": "uvx",
      "args": ["opensearch-mcp-server-py"],
      "env": {
        "OPENSEARCH_URL":            "{{https://collection-id.region.aoss.amazonaws.com}}",
        "AWS_OPENSEARCH_SERVERLESS": "true",
        "AWS_REGION":                "{{us-east-1}}",
        "AWS_PROFILE":               "{{your-aws-profile}}"
      }
    }
  }
}
```

You can also add the server globally via the CLI so it is available in every project:

```
claude mcp add opensearch \
  --command uvx \
  --args opensearch-mcp-server-py \
  --env OPENSEARCH_URL={{https://your-domain-endpoint.region.es.amazonaws.com}} \
  --env AWS_REGION={{us-east-1}} \
  --env AWS_PROFILE={{your-aws-profile}}
```

## Claude Desktop
<a name="mcp-server-configure-claude-desktop"></a>

Open **Settings > Developer** and edit `claude_desktop_config.json`:

```
{
  "mcpServers": {
    "opensearch": {
      "command": "uvx",
      "args": ["opensearch-mcp-server-py"],
      "env": {
        "OPENSEARCH_URL": "{{https://your-domain-endpoint}}",
        "AWS_REGION": "{{us-east-1}}",
        "AWS_PROFILE": "{{your-aws-profile}}"
      }
    }
  }
}
```

## Kiro CLI
<a name="mcp-server-configure-kiro-cli"></a>

If you use the Kiro CLI, add the server to your project's `.kiro/settings/mcp.json` using the same JSON structure shown above. The Kiro CLI picks up the config automatically when you start a session in that directory.