> For the complete documentation index, see [llms.txt](https://docs.arcade.software/kb/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.arcade.software/kb/interactive-demo/leverage/mcp.md). # MCP ### What it does The Arcade MCP server connects AI tools — Claude, ChatGPT, Gemini, and others — to your Arcade workspace, so you can ask questions about your interactive demos' performance in natural language instead of opening Insights. {% hint style="info" %} Want to generate AI videos through Claude, ChatGPT, Gemini, or another AI tool instead? That's the same MCP server, with its own tools for that — see AI Video's [MCP](https://docs.arcade.software/kb/ai-video/mcp/) section for setup and what you can build. {% endhint %} ### Prerequisites An Arcade account with a workspace ([sign up at app.arcade.software](https://app.arcade.software)). ### How do I connect my AI tool? Connecting works the same way no matter which tools you plan to use afterward — see AI Video's [Supported Clients](https://docs.arcade.software/kb/ai-video/mcp/supported-clients/) page for exact per-client setup steps (Claude Web, Claude Desktop, Claude Code, ChatGPT, Gemini, Cursor, and others). ### Do I need a paid plan? The insights tools require a Growth plan or above. On Free, your AI tool can still browse your demo list, but performance, company, trending, and attribution questions will return a plan error. ### How do permissions work? When connecting Arcade, you'll be prompted to grant permissions. You can choose to: * Allow once * Allow always * Require approval Permissions are grouped into read actions (such as insights) and write actions. Delete actions are not currently supported. *** ## Security ### Does Arcade train on my data? No. Arcade does not train on your data. ### What happens when I connect Arcade to third-party tools like Claude? When you connect Arcade to an external AI tool, data shared through that workflow becomes subject to the third party's policies. This is especially important when using free AI plans or working with sensitive information. ### Are free AI tools safe for sensitive data? Free AI tools may store inputs, use data to improve their systems, or have different retention policies. For sensitive use cases, consider using paid or enterprise AI plans and review the provider's data policies. ### How does data access actually work? Claude does not directly browse your Arcade dashboard. Instead, Claude calls specific MCP tools, and those tools return scoped data based on the request — this creates more controlled access. ### Can I control what the MCP can do? Yes. You can control permissions by requiring approval, restricting actions, or limiting access to specific capabilities. ### What should enterprise teams consider? For sensitive use cases: use enterprise AI plans, review third-party security policies, limit permissions where needed, and follow your internal governance standards. ### Can Claude or another AI tool see what's inside my interactive demo? Not through Arcade's MCP connector today — the MCP tools only return performance data (plays, completions, CTA clicks, step drop-off, viewing companies) and don't expose a demo's actual screenshots, hotspot text, or captions. That said, most capable AI assistants — including Claude — can review a demo's actual content a different way: if you give the assistant your demo's share link directly, and it has its own web-browsing capability, it can open the page, step through it, and read what's rendered — screenshots, hotspot and callout text, captions, and the overall flow. This works because the assistant is browsing the published page itself, not calling an Arcade API. {% hint style="info" %} This depends on the AI tool having its own browsing capability — it isn't something Arcade's MCP server provides. Review quality also varies by tool and how well it renders JavaScript-heavy pages. {% endhint %} *** ## Limitations * Insights tools require a Growth plan or above * Cannot create or edit interactive demos through MCP * ChatGPT gets two extra lookup tools (`search` and `fetch`) to find a demo by name or keyword — a ChatGPT-specific requirement, not a capability other clients are missing * Behavior may vary depending on the AI client being used ## Troubleshooting | Issue | Solution | | -------------------------------------------------------- | ------------------------------------------------------------------------------ | | "Only Growth and above plans can request insights" | Your workspace plan doesn't include insights data. Upgrade to Growth or above. | | No data found for a demo | Double-check the demo ID — use `list_demos` to find it. | | Insights takes too long, or times out | Ask again — this is usually a transient backend timeout. | | "This operation requires the arcade:insights scope" | Reconnect the MCP server — your authorization may not include insights access. | | Trending, embedded, or review-and-refresh questions fail | Those specific views need a signed-in connection, not an API key. | --- # Agent Instructions This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com. ## Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation by asking a question. Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters: ``` GET https://docs.arcade.software/kb/interactive-demo/leverage/mcp.md?ask=&goal= ``` `ask` is the immediate question: it should be specific, self-contained, and written in natural language. `goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case. 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.