The full upstream README, mirrored here for reference. Install config, tool schemas, adoption signals, and an original overview live on the Amazon Ads listing page.
Connect Claude, ChatGPT, Cursor, or any MCP client to your live Amazon Advertising data — and, with approval guardrails, let it act: negate wasted search terms, adjust budgets and bids, launch campaigns.
Remote server. One URL, both transports (streamable HTTP and SSE). No API keys. Free on every plan.
Used by 2,000+ brands and agencies in 20+ countries · $1B+ ad spend optimized · 2026 Amazon Ads Partner Awards — Technology Innovation Global Finalist.
Add a remote MCP server with the URL:
The client opens a browser window for the OAuth sign-in on first use. In Claude Code:
Add SellerMate as a custom connector using the same URL. (Connector support varies by ChatGPT plan; see the setup guide.)
Full walkthroughs per client: docs/getting-started.md
| You ask | It uses |
|---|---|
| "Why did ACOS jump last week?" | campaign / targeting / placement performance |
| "Which search terms burn budget with zero orders?" | get_search_term_performance |
| "Negate those" (queued for approval) | add_negative_keywords |
| "Shift budget to the winners" | update_campaign — policy-checked |
| "What's my true TACoS this month?" | retail + ads performance |
| "Launch a campaign for my new product" | create_campaign |
| "How do my Search Query Performance trends look?" | get_sqp_report |
| "Set up dayparting so we stop overspending at night" | create_dayparting |
The server exposes 50+ tools across reporting, campaign management, targeting, negatives, automations, dayparting, DSP, vendor/retail analytics, and account organization. The complete reference: docs/capabilities.md.
The AI proposes; your policies decide. Every write action is checked server-side against workspace guardrails your admin controls — minimum/maximum budgets and bids, change-multiplier caps relative to current values, per-currency floors. Sensitive changes can require explicit admin approval before anything touches Amazon, and every action lands in an audit log. The model cannot exceed the limits you set, because the platform enforces them on every call — not the prompt.
Details: docs/security.md
Amazon's Ads MCP server (open beta) is a strong option if you hold Amazon Ads API partner credentials and are building your own integration. SellerMate's MCP is built for sellers and agencies who want the same conversational access without developer setup, plus a managed guardrail layer:
| SellerMate MCP | Amazon Ads MCP (open beta) | |
|---|---|---|
| Requirements | Free SellerMate account | Active Amazon Ads API partner credentials |
| Setup | One click from Claude's directory, or one URL | Developer OAuth (Login with Amazon) configuration |
| Write safety | Policy guardrails + approval queue + audit trail, enforced server-side | Bring your own guardrails |
| Data scope | Ads + Search Query Performance, retail, vendor & inventory context | Ads API surface (campaigns, reporting, account, billing) |
| Cost | Free (read/analysis); write actions on paid plan | Free API access for approved partners |
They also work side by side. Longer discussion: docs/vs-amazon-official-mcp.md
Is it really free? Yes — connecting and analyzing your data is free, no credit card. Write actions and SellerMate's purpose-built AI agent are on the paid plan.
Do I need a paid Claude account? No. Free Claude works.
Is my data safe? OAuth 2.1 + PKCE, zero caching, account-level access control, read-only by default, full audit log. See docs/security.md.
Which clients work? Claude.ai, Claude Desktop, Claude Code, ChatGPT (custom connector), Cursor, VS Code, Windsurf, Gemini CLI — anything that speaks MCP over streamable HTTP or SSE; the endpoint serves both.
Where does the data come from? Your own Amazon Advertising account, connected to SellerMate. Nothing is scraped; nothing is shared across accounts.
SellerMate.AI — AI-powered Amazon advertising automation. © Sellermate Technologies Pvt Ltd.