# Connect to an AI Client

The MCP server works with any AI client that supports the Model Context Protocol. Select your client for step-by-step setup instructions.

<table data-view="cards"><thead><tr><th></th><th></th><th data-hidden data-card-target data-type="content-ref"></th></tr></thead><tbody><tr><td><mark style="color:blue;"><strong>Claude Code</strong></mark></td><td>Anthropic's CLI. Supports project-scoped <code>.mcp.json</code> for team sharing.</td><td><a href="/pages/qWB407lbsVjG0zDw1T84">/pages/qWB407lbsVjG0zDw1T84</a></td></tr><tr><td><mark style="color:blue;"><strong>Claude Desktop</strong></mark></td><td>Anthropic's desktop AI app. Standard <code>mcpServers</code> config.</td><td><a href="/pages/BRRlG7yfh23mQm3t4RJT">/pages/BRRlG7yfh23mQm3t4RJT</a></td></tr><tr><td><mark style="color:blue;"><strong>Cline</strong></mark></td><td>Open-source VS Code AI agent with MCP marketplace. Uses a separate <code>cline_mcp_settings.json</code>.</td><td><a href="/pages/thuMdpVouFfpCm3r5IGJ">/pages/thuMdpVouFfpCm3r5IGJ</a></td></tr><tr><td><mark style="color:blue;"><strong>Continue</strong></mark></td><td>Open-source extension for VS Code and JetBrains. Agent mode only.</td><td><a href="/pages/gbYHSai4YYKETl55xX2v">/pages/gbYHSai4YYKETl55xX2v</a></td></tr><tr><td><mark style="color:blue;"><strong>Cursor</strong></mark></td><td>AI-native IDE. Config via <code>~/.cursor/mcp.json</code>.</td><td><a href="/pages/xH6Dkm5BQQlhcnWVfGCI">/pages/xH6Dkm5BQQlhcnWVfGCI</a></td></tr><tr><td><mark style="color:blue;"><strong>Gemini CLI</strong></mark></td><td>Google's open-source terminal agent. Standard <code>mcpServers</code> config.</td><td><a href="/pages/SWIFpnFgo7LLcJkoBfvl">/pages/SWIFpnFgo7LLcJkoBfvl</a></td></tr><tr><td><mark style="color:blue;"><strong>JetBrains AI Assistant</strong></mark></td><td>Built into IntelliJ-platform IDEs since 2025.2. Configured via IDE settings UI.</td><td><a href="/pages/SPLtFdy2fktg0xl4LRuD">/pages/SPLtFdy2fktg0xl4LRuD</a></td></tr><tr><td><mark style="color:blue;"><strong>Kiro</strong></mark></td><td>AWS AI IDE for spec-driven development. Config via <code>.kiro/settings/mcp.json</code>.</td><td><a href="/pages/4PwBpzf4wlVMJCyAH9Hc">/pages/4PwBpzf4wlVMJCyAH9Hc</a></td></tr><tr><td><mark style="color:blue;"><strong>OpenCode</strong></mark></td><td>Open-source terminal AI agent supporting 75+ models. Uses <code>opencode.json</code> with <code>type: local</code>.</td><td><a href="/pages/hmnwv8h1djAtALiJ65DE">/pages/hmnwv8h1djAtALiJ65DE</a></td></tr><tr><td><mark style="color:blue;"><strong>VS Code</strong></mark></td><td>GitHub Copilot agent mode. Config via <code>.vscode/mcp.json</code>.</td><td><a href="/pages/MX9outgr6lGdU8wjv1x7">/pages/MX9outgr6lGdU8wjv1x7</a></td></tr><tr><td><mark style="color:blue;"><strong>Windsurf</strong></mark></td><td>AI-native IDE by Codeium. Note: 100-tool limit applies.</td><td><a href="/pages/Cjw3LF8e4ncc4bkcKVeJ">/pages/Cjw3LF8e4ncc4bkcKVeJ</a></td></tr><tr><td><mark style="color:blue;"><strong>Zed</strong></mark></td><td>High-performance code editor. Uses <code>context_servers</code> key.</td><td><a href="/pages/OQ0t5Wg1gh3Od1TDQcma">/pages/OQ0t5Wg1gh3Od1TDQcma</a></td></tr><tr><td><mark style="color:blue;"><strong>Other MCP-Compatible Clients</strong></mark></td><td>Generic stdio config, plus notes on ChatGPT, n8n, and OpenAI Codex CLI.</td><td><a href="/pages/Es5aSc7Hr6Wjr3fa8kKR">/pages/Es5aSc7Hr6Wjr3fa8kKR</a></td></tr></tbody></table>

## Configuration reference

Most clients use the same `mcpServers` JSON structure. Point the client at the `ionoscloud-mcp` binary and pass your credentials as environment variables:

```json
{
  "mcpServers": {
    "ionoscloud": {
      "command": "/path/to/ionoscloud-mcp",
      "env": {
        "IONOS_TOKEN": "your-api-token"
      }
    }
  }
}
```

**Three clients differ from this pattern:** [<mark style="color:blue;">Zed</mark>](/cloud/ai/mcp-server/connect-to-an-ai-client/zed.md) uses a `context_servers` key with a different JSON structure. [<mark style="color:blue;">OpenCode</mark>](/cloud/ai/mcp-server/connect-to-an-ai-client/opencode.md) uses a `"type": "local"` entry with a `command` array and an `environment` key instead of `env`. [<mark style="color:blue;">JetBrains AI Assistant</mark>](/cloud/ai/mcp-server/connect-to-an-ai-client/jetbrains.md) is configured entirely through the IDE settings UI; no JSON file to edit.

For Object Storage tools, also add `IONOS_S3_ACCESS_KEY` and `IONOS_S3_SECRET_KEY`. For more information, see [<mark style="color:blue;">Authentication</mark>](/cloud/ai/mcp-server/configuration/authentication.md).

## Add the IONOS CLOUD Documentation MCP

The <code class="expression">space.vars.ionos\_cloud</code> documentation site exposes a free, public MCP server at `https://docs.ionos.com/cloud/~gitbook/mcp`. Add it alongside the infrastructure server to give your AI assistant access to both live account data and product documentation in the same session:

```json
{
  "mcpServers": {
    "ionoscloud-docs": {
      "url": "https://docs.ionos.com/cloud/~gitbook/mcp"
    },
    "ionoscloud": {
      "command": "/path/to/ionoscloud-mcp",
      "env": {
        "IONOS_TOKEN": "your-api-token"
      }
    }
  }
}
```

No credentials are required for the documentation server. The two servers are completely independent; use either one alone or both together.

Mixing `url` (remote, HTTP) and `command` (local, stdio) entries in the same `mcpServers` block is valid. Each entry uses its own transport; the AI client merges their tool catalogs automatically.

For more information, see [<mark style="color:blue;">Composing Multiple MCP Servers</mark>](/cloud/ai/mcp-server/use-cases/agentic-workflows.md#composing-multiple-mcp-servers) for example prompts.


---

# Agent Instructions: Querying This Documentation

If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://docs.ionos.com/cloud/ai/mcp-server/connect-to-an-ai-client.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.