Global Finalist2026 Amazon Ads Partner Awards
Listed inClaude’s MCP Directoryclaude.ai

FREE FOR ALL USERS

Your Amazon Ads data, in the AI you already use

Ask questions in plain English, get answers from your live Amazon data — and when you're ready, let the AI act for you, inside guardrails your admin controls.

Works with Claude · ChatGPT · Gemini CLI · Cursor · VS Code · Windsurf

Write a message…

HOW IT WORKS

From this week's questions to this week's decisions

Your team already asks these questions. Now the AI they use can answer them from your live Amazon data — and act on the answers.

Traffic & Performance

  • Why did ACOS jump last week?
  • Which campaigns win top-of-search?

Wasted Spend

  • Which search terms burn budget with zero orders?
  • Which keywords should I negate?

Growth & Retail

  • What's my true TACoS this month?
  • Which new products should get a campaign?

SellerMate MCP

One URL · read + write · policy-checked

Claude

Conversational deep analysis

Claude Desktop

Native app + your local files

Claude Code

Terminal & automated workflows

ChatGPT

Custom connector

Cursor

In-editor data access

Decisions out

  • Root cause for the ACOS spike
  • Negation list applied with approval
  • Budget shifts on proven winners
  • New campaign live on Amazon

Real-time · policy-checked

USE CASES

Ask it anything. Then act on it.

Five things sellers do with SellerMate's MCP every day — exactly as they happen in Claude.

Claude + SellerMateMCP · api.sellermate.ai/mcp/sse
Which search terms are wasting my budget?

Pulled 30 days of search-term data across 12 active campaigns. 8 terms spent $842 with zero orders — the top three:

  • 1knife set premium$212 · 0 orders314 clicks
  • 2kitchen gadgets cool$147 · 0 orders203 clicks
  • 3chef knife japanese steel$119 · 0 orders96 clicks

Want me to add these as negative exact keywords? I can queue them for approval.

get_search_term_performance · 3.8s

Find the leaks in seconds, not spreadsheets.

  • Search-term audits that took an afternoon in bulk files come back in one question.
  • The AI cross-checks spend, clicks, and orders — and flags exactly what to negate.
  • Follow up with "negate them" and the change is queued through your guardrails.

GET CONNECTED

Add SellerMate to Claude in one click

SellerMate is listed in Claude's connector directory, so there is nothing to paste and no API keys to manage. Two steps and you are asking questions about your own Amazon data.

  1. 1

    Click Add to Claude. SellerMate opens in Claude's connector directory.

  2. 2

    Sign in with your free SellerMate account. That is it — start asking.

MCP DirectoryListed in
Claude’s Directory
claude.ai
Add to Claude

Opens in Claude. Sign in with SellerMate and you’re connected.

Using ChatGPT, Gemini CLI, Cursor, VS Code or another MCP client? Copy this URL instead.

https://api.sellermate.ai/mcp/sse
OAuth 2.1 + PKCENo API keysFree on every plan

GUARDRAILS & VISIBILITY

Safe to hand to an AI

Workspace policy guardrails - budget and bid limits

Policy Guardrails Your Admin Controls

Every write action is checked against your workspace policies before it touches Amazon: minimum and maximum budgets and bids, change-multiplier caps relative to current values, and per-currency floors. The AI cannot exceed the limits you set — the platform enforces them on every call.

Admin approval queue and full audit trail for AI actions

Approvals & a Full Audit Trail

Require admin approval for sensitive changes — queued actions don't modify your account until an admin approves them. Every action is logged with who asked, the AI's stated reasoning, the policy checks it passed, and the result. No black-box changes, ever.

Workspace users with per-user MCP connection status and roles

Control who connects, down to the account

Manage exactly who in your workspace can connect an AI to SellerMate. See MCP connection status per user, scope each teammate to specific Amazon accounts, and revoke access in one click. Combined with workspace roles (Owner, Analyst, Client, WhiteLabel Client), you control both who connects and what they can see.

MCP usage dashboard - total calls, success rate, and per-tool activity log

See every call, with a full activity log

Every MCP tool invocation is logged — which tool was called, by which user, against which workspace and Amazon account, how long it took, and whether it succeeded. View total calls, success rate, errors, and average response time at a glance, or drill into a detailed activity log with date-range filters. No black-box AI — you always know exactly what your team is querying.

One URL. Every tool your ads team needs.

Campaigns & Placements

Spend, sales, ACOS, ROAS, CTR, CPC and more for any campaign, any date range — down to placement and hour of day.

get_campaign_performanceget_placement_performanceget_hourly_performance

Targeting & Keywords

Keyword and product-target performance by match type and ad group. Find winners, cut waste, add new targets.

get_targeting_performancelist_targetsadd_targets

Search Terms & SQP

The queries shoppers actually typed, plus Amazon Search Query Performance — organic vs paid at the query level.

get_search_term_performanceget_sqp_report

Products & Catalog

Per-ASIN ad performance across Sponsored Products, Brands, and Display — plus a browsable product catalog.

get_advertised_product_performancelist_product_catalog

Orders, Inventory & Vendor

Order-level reports, stock levels, FBA returns, and Vendor Central shipped revenue, COGS, and net PPM.

get_orders_reportget_inventoryget_vendor_sales_performance

Write Actions

Create campaigns, update budgets, bids, and targets, add negatives — all policy-checked, all logged.

create_campaignupdate_campaignupdate_targetadd_negative_keywords

Knowledge & Memory

Upload brand guides, SOPs, pricing sheets, or data exports — the AI pulls them in whenever they're relevant.

list_knowledge_and_memory_files

Account Context & Deep Links

Discovers your workspaces and accounts, and deep-links you into 25+ SellerMate features beyond the MCP tools.

get_user_contextresolve_accountget_sellermate_capabilities

Enterprise-grade security

OAuth 2.1 + PKCE

Industry-standard secure authentication — the same protocol used by banking apps. Your credentials never touch the AI client.

Firebase Authentication

Sign in with Google or email/password through Firebase Auth. Multi-factor authentication supported.

Account-Level Access Control

Two-layer permissions: workspace account access plus per-user MCP grants, with separate read and write levels — checked live on every single tool call.

Zero Data Caching

Your advertising data flows through in real-time. Nothing is stored or cached by the AI client. 28-day session persistence with JWT tokens.

How does this compare to Amazon's official MCP?

Amazon released its own Amazon Ads MCP server in open beta — availability is scoped to Amazon Ads partners with active API credentials, while SellerMate's MCP connects any free SellerMate account with server-enforced write guardrails. Read the full SellerMate MCP vs Amazon Ads MCP comparison.

SellerMate MCP: answers before you connect

FREQUENTLY ASKED QUESTIONS

Claude (web, desktop, and mobile), Claude Code, ChatGPT, Cursor — and any other MCP-compatible client. It's one URL: https://api.sellermate.ai/mcp/sse. No API keys, no plugins to install.

Campaign, targeting, search-term, SQP, placement, and advertised-product performance; orders, inventory, and FBA returns; plus Seller Central and Vendor Central retail metrics — across every marketplace and account you've connected to SellerMate.

Yes — if you grant write access. It can launch Sponsored Products campaigns, update budgets and bids, and add targets or negatives. Every write is checked against workspace policies your admin sets, can require approval before it applies, and is logged with the AI's stated reasoning.

Authentication uses OAuth 2.1 with PKCE — your credentials never touch the AI client. Access is checked per user, per account, on every single call, with separate read and write levels. Nothing is cached: data flows through in real time, and admins can revoke a user's access in one click.

Nothing extra. MCP access is included in every SellerMate plan — including the free tier.

An open standard that lets AI assistants securely connect to tools and data. SellerMate hosts a remote MCP server, so there's nothing to install or run yourself — and because SellerMate is listed in Claude's connector directory, adding it takes a single click.

Amazon's Ads MCP server (open beta) is an API surface for Amazon Ads partners with active API credentials — great for developers building custom agents. SellerMate's MCP is a managed operator for sellers and agencies: you connect with a free SellerMate account (no API credentials), and every write action is checked against guardrails your admin controls, with approval queues and a full audit trail. You can run both side by side — see the full comparison at sellermate.ai/product/mcp-vs-amazon-mcp.

Click Add to Claude above. SellerMate is listed in Claude's connector directory, so it opens there and you simply sign in with your free SellerMate account — secure OAuth, no API keys, nothing to paste. Then open a new chat and ask something like "Show me my campaign performance for the last 30 days." Using ChatGPT, Gemini CLI, Cursor, VS Code, Windsurf or another MCP client? Copy the server URL instead and add it as a custom connector.

Ready to talk to your Amazon Ads data?

Connect SellerMate to Claude in under 2 minutes. Free for all users — no credit card required.

Free for all SellerMate users including free tierNo API keys, no coding, no dashboard switchingWorks with Claude, ChatGPT, Gemini CLI, Cursor and VS Code
Book Demo