Integration MCP Server

The Magify Integration MCP server gives your AI coding assistant direct access to the Magify documentation and the public SDK API. Instead of guessing from stale training data, the assistant looks up the current integration guide, finds the right page, and checks real method signatures before it answers or writes code.

The server is public, read-only, and needs no sign-in or API key. It returns documentation excerpts and public SDK signatures from the latest releases. Your own assistant composes the answer.

When to use it

Integrating the Magify SDK into an Android, iOS, Flutter, or Unity project and you want step-by-step help in your editor.
Checking the exact name, parameters, or return type of an SDK method without leaving your code.
Finding the right documentation page for a feature, such as purchases, attribution, or campaigns.

Supported platforms: android, ios, flutter, unity, and general (documentation that is not tied to one platform).

Server details

Property	Value
Transport	Streamable HTTP (remote)
Authentication	None
Access	Read-only

The server URL is:

https://help.magify.com/mcp

Add it to an MCP client as shown below; it is not a web page to open in a browser.

Connect your client

Pick the client you use. All of them need a version that supports remote (HTTP) MCP servers.

Claude Code

Run once in your terminal:

claude mcp add --transport http magify-help https://help.magify.com/mcp

By default the server is available only to you in the current project. Add --scope user to make it available in every project, or --scope project to save it to .mcp.json and share it with your team.

Check the connection with claude mcp list, or run /mcp inside a Claude Code session.

If you edit the JSON configuration by hand, the entry needs an explicit type:

{
  "mcpServers": {
    "magify-help": {
      "type": "http",
      "url": "https://help.magify.com/mcp"
    }
  }
}

Claude Desktop

Remote servers are added as custom connectors, not through claude_desktop_config.json (that file is for local servers).

Open Customize > Connectors.
Click + Add, then Add custom connector.
Enter the name magify-help and the server URL from Server details, then click Continue.
Under Authentication, choose No sign in, then click Add.
In a conversation, click + in the lower left, open Connectors, and make sure magify-help is switched on.

On Team and Enterprise plans an Owner adds the connector first, in Organization settings > Connectors; members then connect to it from Customize > Connectors.

Codex

The Codex CLI, the Codex IDE extension, and the ChatGPT desktop app share one MCP configuration, so you only need to add the server once.

Run in your terminal:

codex mcp add magify-help --url https://help.magify.com/mcp

Or add it to ~/.codex/config.toml (all projects) or .codex/config.toml (current trusted project):

[mcp_servers.magify-help]
url = "https://help.magify.com/mcp"

Check the connection with codex mcp list, or run /mcp inside a Codex session. In the IDE extension, open the gear menu, select MCP servers > Add server, choose Streamable HTTP, enter the server URL, save, and restart the extension.

Cursor

Add the server to ~/.cursor/mcp.json (all projects) or .cursor/mcp.json (current project):

{
  "mcpServers": {
    "magify-help": {
      "url": "https://help.magify.com/mcp"
    }
  }
}

You can also manage servers from the Customize page in Cursor. After saving, the server appears in the MCP list and its tools become available to the agent.

VS Code (GitHub Copilot)

Run MCP: Add Server from the Command Palette, choose HTTP, and enter the server URL. Or create .vscode/mcp.json in your workspace:

{
  "servers": {
    "magify-help": {
      "type": "http",
      "url": "https://help.magify.com/mcp"
    }
  }
}

Open Copilot Chat in Agent mode, click Configure Tools, and check that the magify-help tools are enabled. VS Code also accepts a workspace-root .mcp.json that uses a top-level mcpServers object instead of servers.

Unity Editor (AI Assistant)

Unity AI Assistant (package com.unity.ai.assistant 2.x) can call remote MCP servers right inside the Unity Editor.

Open Edit > Project Settings > AI > Assistant MCP Extensions.
In the Server Configuration section, click Open. The configuration is stored in UserSettings/mcp.json of your project.
Enable MCP and add the server:

{
  "enabled": true,
  "mcpServers": {
    "magify-help": {
      "type": "http",
      "url": "https://help.magify.com/mcp"
    }
  }
}

Save the file and click Refresh File and Servers, then select magify-help and click Inspect to check that its tools are listed.

MCP support in Unity AI Assistant is in preview; menu names may differ between package versions. See the Unity AI Assistant MCP documentation for the current details. If you write Unity code in Rider or VS Code, you can also connect the server there (see the JetBrains IDEs and VS Code sections).

Gemini CLI

Run once in your terminal:

gemini mcp add --transport http magify-help https://help.magify.com/mcp

By default the server is saved to the current project (.gemini/settings.json). Add --scope user to save it to ~/.gemini/settings.json and use it in every project.

If you edit settings.json by hand, Gemini CLI uses httpUrl (not url) for Streamable HTTP; url is reserved for the older SSE transport:

{
  "mcpServers": {
    "magify-help": {
      "httpUrl": "https://help.magify.com/mcp"
    }
  }
}

Check the connection with gemini mcp list, or run /mcp inside a Gemini CLI session.

Windsurf

Windsurf is now documented as Devin Desktop, and it has two agents with separate MCP configuration.

For the Cascade agent, open the Actions (...) menu in the top right of the Cascade panel, click Open MCP config file in the MCPs section, and add the server. Remote HTTP servers use serverUrl:

{
  "mcpServers": {
    "magify-help": {
      "serverUrl": "https://help.magify.com/mcp"
    }
  }
}

For the Devin Local agent, which is the default for new tabs, use the Devin CLI configuration instead:

devin mcp add magify-help https://help.magify.com/mcp

Check the connection with devin mcp list. For Cascade, the server appears in the MCPs section of the same Actions menu, where you can switch it on or off.

JetBrains IDEs

JetBrains AI Assistant supports remote MCP servers over Streamable HTTP. This section covers AI Assistant only.

Open Settings > Tools > AI Assistant > Model Context Protocol (MCP).
Click Add and choose the JSON configuration option.
Paste the entry below, choose the server level (global or current project), and click OK, then Apply.

{
  "mcpServers": {
    "magify-help": {
      "url": "https://help.magify.com/mcp"
    }
  }
}

Check that the Status column shows a successful connection. Click the icon in that column to see the server's tools.

Zed

Open Settings > AI > MCP Servers, click Add Server, and choose Add Remote Server. Or add the server to your Zed settings file (open it with the zed: open settings file command). Zed calls MCP servers context_servers:

{
  "context_servers": {
    "magify-help": {
      "url": "https://help.magify.com/mcp"
    }
  }
}

Open Settings > AI > MCP Servers and check the indicator dot next to magify-help: green with the tooltip "Server is active" means it is connected. In the Agent Panel, mention the server by name to help the model pick its tools.

Other MCP clients

Any client that supports remote MCP servers over Streamable HTTP can use this server. Enter these values in its MCP settings:

Setting	Value
Name	`magify-help`
Transport	Streamable HTTP (sometimes labeled `http`)
URL	the server URL from Server details
Authentication	None

Field names vary by client: the URL is usually url, and some clients use serverUrl or httpUrl. Check your client's documentation for the exact format.

Clients that run only local stdio servers cannot connect to this server directly.

Available tools

You do not call these yourself: the assistant picks the right one from your question.

get_integration_guide — the step-by-step integration guide for a platform, or for one topic within it (for example, purchases or AppsFlyer). Arguments: platform, optional topic.
search_docs — documentation excerpts that match a question, with links to the source pages. Arguments: query, optional platform and limit (up to 10).
get_page — the full content of a specific documentation page. Arguments: url.
get_sdk_api — public SDK signatures from the latest release, or the details of one class or method. Arguments: platform, optional symbol.

Example prompts

Ask in plain language and name the platform, so the assistant can scope the lookup:

How do I initialize the Magify SDK on iOS?
How do I connect AppsFlyer in Unity?
What is the signature of the purchase method on Android?

Troubleshooting

The server does not connect or its tools do not appear. The client must support remote HTTP MCP servers (Streamable HTTP). Older clients, and some tools that only run local stdio servers, cannot use it. Update the client to the latest version.

Claude Code reports that the entry has a url but no type. Add "type": "http" to the JSON entry, or add the server with the claude mcp add --transport http command shown above.

The tools are connected but the assistant does not use them. Mention Magify and the platform in your prompt, and check that the server is enabled for the current chat or agent session.

The connection times out. Claude Code, Cursor, and VS Code connect from your machine, so your network or proxy must allow HTTPS requests to help.magify.com. Claude Desktop connects through Anthropic's cloud, so the restriction does not apply there.