Amazon Ads MCP server vs Xnurta MCP: what is the difference?

What is the Amazon Ads MCP server, according to Amazon?

It is Amazon's own MCP server, and Amazon describes it as a translation layer that turns natural language prompts into structured Amazon Ads API calls. Once an AI agent connects, Amazon says it can create, update or delete campaigns, run reporting queries, manage account-level settings and reach billing and financial data.

Amazon also ships tools on the server. A tool bundles several API actions into one workflow. The example Amazon gives is a Sponsored Products tool that builds a full campaign in a single workflow, work that would normally take three or more separate operations, and hands back a campaign that only needs review and approval. Amazon names custom-built agents, Claude, ChatGPT and Gemini as clients that can connect.

On access, Amazon's announcement says the server is available to Amazon Ads partners with active API credentials. Check Amazon's own MCP overview documentation for the current terms, setup and tool list. This guide reports only what Amazon has published.

What does Xnurta MCP give your AI to work with?

It gives your assistant Xnurta's data and Xnurta's controls, behind your own Xnurta login. Xnurta MCP connects Claude, ChatGPT and other MCP clients to your Xnurta data, at https://mcp.xnurta.com/mcp.

Three things set it apart, and each one is about control.

It reads from Xnurta data. Three query tools cover ads performance, entity metadata and the operation log. Queries are read only. You get about 15 months of history across 22 marketplaces, daily data by default and hourly data for SP, SB and SD campaigns and SP keywords. Performance can be broken out by audience.

It manages AI managed groups. A managed group holds campaigns that Xnurta's AI manages. Through MCP you can create, edit, schedule and switch groups on or off, one at a time or in bulk. You can set the name, goal, target ACOS, total budget, AI personality, campaign membership, schedule and whether each part of the action space is on. Brand, non-brand and competitor mode, and every word-list feature, stay in the Xnurta platform.

It reads the operation log. The operation log holds human and AI actions on campaigns, ad groups, targets, bids, budgets and managed groups. Filter it by operator: the AI, a human, an automation rule or one person's email.

How do the two servers compare side by side?

The Amazon column reports what Amazon states publicly. The Xnurta column reports what Xnurta MCP does. Where Amazon's public material is silent, the row stays out.

What happens before a change goes live?

With Xnurta MCP, anything involving money, bids or archiving returns a preview first. The preview names the objects and shows each change from the old value to the new one. Nothing is submitted until you confirm.

That rule covers a lot of daily work: a budget raised on three campaigns, bids lowered on a keyword list, a non-converting ASIN archived. Ask the assistant to list the exact values it plans to change, and read them before you say yes. Archiving cannot be undone, so pause a campaign if you only want it to stop.

For Amazon's server, Amazon describes one review step: the Sponsored Products tool returns a campaign that only needs review and approval. For anything beyond that, read Amazon's documentation. This guide will not guess.

How do you see who changed a campaign?

On Xnurta MCP, you ask the operation log. It is the answer to "was that us, the AI or an automation rule?" and it is read only, so asking changes nothing.

In ad account [account name], who changed campaign [campaign name or ID]'s budget, and when? List all human actions from the last [30] days.

Call get_user_authorized_context and list the ad accounts I'm authorized to access.

Who changed my Amazon campaign? walks through the filters in detail.

What does a managed-group change look like?

You describe the group, and the assistant shows you the whole configuration before it submits anything. Managed group management modifies your account, so ask for the objects, target ACOS, AI status, action space and campaign list first.

In ad account [account name], create an AI managed group from these SP campaigns: [campaign names]. Set target ACOS to [X]%, leave AI switched off for now, and show me the objects, target ACOS, AI status, action space and campaign list before you submit.

Deleting a managed group archives it, and that is irreversible. Switching a group's AI on or off is a separate setting from deleting it.

Can the AI change anything on Xnurta MCP?

Only if write access is turned on for your account. Reading never changes anything. Writing comes in three kinds of write, each with its own permission scope: managed group management, ad-entity editing and Sponsored Products campaign creation. A scope that edits ads does not necessarily create campaigns.

Read-only vs read-write MCP for Amazon advertisers explains which access to start with.

Which one fits your team?

Pick by what your team already holds and what you want the assistant to do. Amazon's server is aimed at Amazon Ads partners with API credentials. Xnurta MCP is aimed at people who already run their ads in Xnurta.

How do you try Xnurta MCP?

Add https://mcp.xnurta.com/mcp as a custom connector in Claude, sign in with your Xnurta login, then ask the assistant to install the skills from the Xnurta repository on GitHub. It is a few steps and you write no code. Connect Claude or ChatGPT to Xnurta MCP gives the exact screens for Claude, ChatGPT, Codex and tokens.

Start with the read-only prompts above. When you are ready to compare approaches in more depth, How to choose an Amazon Ads MCP lists what to check in any MCP for Amazon Ads, and What is an Amazon Ads MCP? covers the basics.