Then let the Badger dig. The Ad Badger Amazon Ads MCP server is the fastest way to query your Amazon PPC data in natural language. Powered by Ad Badger’s intelligent data engine and the Model Context Protocol (MCP).
1-Minute Turnkey Setup
Connect via standard OAuth. No Amazon developer accounts, AWS credentials, or API keys needed.
Sub-Second Performance
Queries Ad Badger’s optimized database directly—bypassing slow, rate-limited Amazon API calls.
Your Goals, Not Amazon’s
Analyze campaigns with true profit and ACOS-focused logic from a team trusted by sellers since 2017.
Read-only by default. It can look at everything and change nothing. Allow changes and it can adjust bids, placement modifiers and budgets, and pause or enable campaigns, keywords, targets and product ads, asking you before every change. It shipped on our public roadmap: more than 240 shipped updates, and never a price increase for existing customers since 2017.
Works on every Ad Badger plan — free trials included.
Why Ad Badger MCP
Built for Seller Profitability, Not Raw API Complexities
The Ad Badger Amazon Ads MCP server connects your Amazon Ads data to Claude, ChatGPT, and other AI assistants. MCP (Model Context Protocol) is an open standard that lets AI apps securely talk to tools like Ad Badger. You ask in plain English; the AI pulls live answers from your account.
50–80%
of ad spend in the average account we audit goes to non-converting targets. Honey badgers famously don’t care. Ours cares about exactly one thing: finding yours. Ask the AI — it answers in one prompt.
Source: Ad Badger platform audits, August 2026.
Amazon’s native MCP beta requires complex AWS developer approvals, custom LWA applications, and suffers from slow response times. Worse, Amazon’s default tools are built to encourage ad volume—not protect your profit margins.
Ad Badger bridges the gap by acting as an intelligent, sub-second reasoning layer between your favorite AI assistant and your ad accounts.
| Amazon Native MCP Beta | Ad Badger MCP Connection |
|---|---|
| ✘ 30–60 min AWS developer setup & LWA approvals | ✔ 1-click OAuth login (takes under 60 seconds) |
| ✘ Incredibly slow response times on large accounts | ✔ Sub-second query speeds from Ad Badger database |
| ✘ Default goal: Maximize Amazon ad spend | ✔ Aligned with YOUR goals: Lower ACOS & wasted spend |
| ✘ Risk of accidental unauthorized changes | ✔ Read-only by default. Changes only if you allow them, and it asks before each one |
The Badger is unbothered by rate limits. Your wasted spend, however, should be very worried.
The Tools
What It Can Do Today
As of our November 2026 release, Ad Badger MCP has 24 tools: 12 that read your account and 12 that can change it once you allow changes. Answers return up to 250 rows.
Read: ask anything about your account
Multi-Account Discovery
list-profiles-tool
Instantly list and switch between all connected seller accounts, vendor profiles, and every Amazon Ads marketplace worldwide.
Campaign Diagnostics & Settings
get-campaigns-tool
Impressions, clicks, spend, sales, orders, ACOS and Top of Search impression share across any date range, next to each campaign’s daily budget, bidding strategy, placement adjustments and serving status. Ask “which campaigns are out of budget?” and get a straight answer. See the weekly budget-pacing workflow.
Keyword Performance & Match Types
get-keywords-tool
Keyword-level ACOS and current bids, click bleeders with zero conversions, and a true Exact vs. Phrase vs. Broad comparison across the whole account.
Search Term Harvesting & Discovery
get-search-terms-tool
See the exact customer search queries triggering your ads, detect non-converting terms for future negation, and find new positive targets.
N-Gram Wasted Spend Patterns
get-n-grams-tool
Find the single words and phrases that waste money across every search term in the account, not just the top 250. The fastest way to build a negative list.
Auto & Product Targets
get-targets-tool
Close-match, loose-match, substitutes and complements, ASIN, brand and category targets, and Sponsored Display audiences, with their bids and performance.
Advertised Products
get-product-ads-tool
Which ASINs spend and sell the most, which have a bad ACOS, and which ads aren’t delivering and why: out of stock, not in the Buy Box, and more.
Portfolio Breakdowns
get-portfolios-tool
Spend, sales and ACOS per portfolio, next to each portfolio’s budget and serving status.
Organic Rank
get-organic-ranks-tool
The terms you track in Rank Tracking: today’s exact organic rank on every results page we check, not just the top 10, plus 30-day movement, biggest gainers and losers, which products dropped out of the results, and your tracked keyword allowance.
Placement Performance
get-placements-tool
Top of Search, Rest of Search, Product Pages, Home and Off Amazon performance per campaign, next to your current placement bid adjustments. Compare up to four date ranges side by side in one call.
Search Query Performance
get-search-query-performance-tool
Compare your conversion rate, impression share, click share, cart-add share and purchase share against the whole Amazon market on every search query and ASIN, across up to 13 weeks or 12 months.
Total ACOS & Organic Sales
get-total-acos-tool
Your Seller Central sales next to your ad metrics: Total Sales, Organic Sales, Ad Sales, Ad Spend, ACOS and Total ACOS (TACOS) per child SKU or parent ASIN, with daily, weekly or monthly trends.
Change: only when you allow it
Bid Changes
set-keyword-bid-tool · set-target-bid-tool
Set new bids on up to 100 keywords or product targets at once, the same bid for all or a different bid for each.
Placement Adjustments
set-placement-adjustments-tool
Set Top of Search, Rest of Search and Product Pages bid adjustments (0% to 900%) on up to 100 Sponsored Products campaigns at once.
Budget Changes
set-campaign-budget-tool
Set the daily budget on one campaign or up to 100 at once.
Pause & Enable
pause-*-tool · enable-*-tool
Switch campaigns, keywords, product targets and product ads off or back on, up to 100 per request. Pausing is always reversible.
🛡️ Changes are off until you turn them on. A new connection is read-only. Once you allow changes, the AI asks before every change, and each one is recorded in your Ad Badger change history as “Badger MCP (AI)”. Changes cover Sponsored Products for now; Sponsored Brands and Sponsored Display stay read-only. Nothing can be deleted. See what shipped on our public roadmap →
Try It
27 Prompts to Try Right Now
Once connected, ask Claude or ChatGPT these exact questions:
Which search terms generated zero sales but spent over $5 in the past week?
Here are 12 search terms with spend over $5 and zero attributed sales in the last 7 days, sorted by wasted spend… [live table from your account]
🌐 Account & Profile Discovery
Start here — see everything the AI can reach before you dig in.
“What Amazon Ads accounts and stores do I have connected in Ad Badger?”
“List all my connected marketplaces and their profile IDs.”
📊 Campaign Performance & Spend Audits
Start wide and let the answer narrow itself, instead of clicking through campaigns one at a time.
“What are my top 5 campaigns by spend over the past 30 days?”
“What was the change in spend over the past 30 days compared to the previous 30 days?”
“Give me a list of all campaigns with an ACOS over 50% in the last 14 days.”
“What was my total campaign spend, sales, and overall ACOS for last month?”
“Which campaigns would you recommend for a budget increase based on ACOS?”
🎯 Keyword Efficiency & Match-Type Analysis
Find the click bleeders. The Badger shows no mercy to keywords that spend and don’t sell.
“What are my top 10 best-performing keywords by sales across all campaigns in the last 30 days?”
“Which keywords in [Campaign Name] have the worst ACOS?”
“Find all keywords that received at least 10 clicks but zero orders in the last 14 days.”
“What performs best across my account: exact match, phrase, or broad?”
“Which keywords have a rising CPC over the last 30 days?”
🔍 Customer Search Term Harvesting
The raw material for your next negative — and your next winner.
“Which search terms generated zero sales but spent over $5 in the past week?”
“Which search terms have a CTR above 1% and at least 3 conversions in the past 7 days?”
“Show me all customer search terms triggering for [Target Keyword] over the last 30 days.”
🧩 Words, Placements, Market Share, TACOS & Rank
Go past keywords: the placements, market benchmarks, organic sales and ASINs behind your spend.
“Which single words across my search terms spent over $25 with no sales in the last 30 days?”
“Compare my Top of Search ACOS for the last 7 days against the previous 7 days.”
“Which search queries do I get clicks on but convert below the Amazon market?”
“What is my Total ACOS and organic sales by product this month compared to last month?”
“How are the close-match and loose-match targets in my auto campaigns performing?”
“Which of my advertised products have the worst ACOS this month, and are any out of stock?”
“Which campaigns are out of budget right now, and what is their Top of Search impression share?”
“Which of my tracked products climbed or dropped the most in organic rank this month, and did any fall out of the results?”
⚡ Make Changes (with changes allowed)
It shows you exactly what it will change and waits for your yes.
“Pause every keyword with at least 20 clicks and zero orders in the last 30 days.”
“Lower bids by 20% on keywords in [Campaign Name] with an ACOS over 60%.”
“Set Top of Search to 35% on [Campaign Name].”
“Raise the daily budget to $50 on the campaigns that ran out of budget this week.”
Field Research
Real Actual Footage of The Badger Using the AI Server to Manage Amazon PPC
Nine years of Amazon PPC research, and this is what we are proudest of. He is checking his search term report.
Amazon sellers and agencies are putting the Ad Badger Amazon Ads MCP to work.
Connect Claude or ChatGPT to your Amazon Ads and ask it what’s working.
As of Oct 30, 2026, the Ad Badger MCP has handled 36,865 calls since launch.
Setup
How to Connect (Under 1 Minute)
No terminal. No code. Sign in with your regular Ad Badger login and approve access. Every new connection starts read-only.
Want these steps open in another tab while you set up? The same walkthrough lives in our help center: How do I connect Ad Badger to Claude or ChatGPT?
Our production endpoint is:
https://ai.adbadger.com/mcp/ads
(Tap the address to select it. Always include the full /mcp/ads path in your configuration.)
Heads up: Claude and ChatGPT change their screens often.
They’re not our products. Anthropic and OpenAI regularly rename menus, move buttons and reorder steps without notice, so what you see may not match the steps below. We last checked them on September 23, 2026. If your screen looks different, go by Anthropic’s guide or OpenAI’s guide, or contact Ad Badger support and we’ll walk you through it. Whatever the menus look like, the job is the same: add a custom connector (Claude) or MCP app (ChatGPT) using the address above, then sign in with your Ad Badger login.
Option A: Claude (Desktop or Web)
Claude is Anthropic’s product and we don’t control its interface — these steps can change at any time. If your screen doesn’t match, Anthropic’s documentation is the authority. Any trouble, contact Ad Badger support and we’ll help.
- Open your profile menu and choose Settings.
- Click Customize at the bottom of the left sidebar, then choose Connectors.
- Click Add in the top right. (Not Discover — that’s Claude’s directory of pre-built connectors, and Ad Badger isn’t listed there.)
- Enter a Name (Ad Badger) and the Remote MCP server URL https://ai.adbadger.com/mcp/ads, then click Continue.
Leave the next screen exactly as you find it. Claude checks our server and pre-selects the right settings — they’re the ones marked Detected. Don’t switch to the OAuth client option marked Recommended (labelled “Use Anthropic’s hosted client metadata” or “Use Claude’s published identity,” depending on your version of Claude). Ad Badger doesn’t use that method, and selecting it will stop the connection from working. The same applies to anything under Advanced — the transport setting is filled in automatically from the address. Just click Add.
Sign in with your standard Ad Badger customer credentials, review the consent screen, and approve. The connection starts read-only.
Done! You can now ask Claude questions about your live Amazon Ads data anytime.
Custom connectors are supported across Claude’s plans. Anthropic occasionally changes which plans include them — check Anthropic’s current documentation if you’re unsure about yours.
On a Team or Enterprise plan? Custom connectors are usually added once for the whole organization by an owner, under Organization settings ➞ Connectors. If you don’t see an Add button, ask whoever administers your Claude workspace.
Option B: ChatGPT
ChatGPT is OpenAI’s product and we don’t control its interface — these steps can change at any time. If your screen doesn’t match, OpenAI’s documentation is the authority. Any trouble, contact Ad Badger support and we’ll help.
- Settings ➞ Security and login ➞ switch on Developer mode.
- Click Plugins in the left sidebar ➞ Create app ➞ Create MCP app.
- Name: Ad Badger. Icon and Description are optional.
- Connection: leave the toggle on Server URL (not Tunnel), enter https://ai.adbadger.com/mcp/ads.
- Authentication: OAuth. Leave Advanced OAuth settings alone — ChatGPT discovers them and sets Registration method to Dynamic Client Registration (DCR) automatically.
- Tick “I understand and want to continue”, then Create.
- On the Add Ad Badger to ChatGPT screen, click Sign in with Ad Badger, sign in, and approve the consent screen. The connection starts read-only.
- In a new chat, type @ and pick Ad Badger.
Two warnings you may see — neither is about Ad Badger. Turning on Developer mode shows “ELEVATED RISK — allows you to add unverified connectors that could modify or erase data permanently.” That is OpenAI’s blanket warning about developer mode itself, not about us. The Ad Badger connection is read-only unless you allow changes, and even then it can only change bids, placement adjustments, budgets and on/off status, after asking you. It cannot erase anything. And while configuring, ChatGPT may say “CIMD is unavailable because the server did not advertise CIMD support.” Expected and harmless: CIMD is a sign-in method Ad Badger doesn’t use, and ChatGPT has already selected the right one.
ChatGPT requires Developer mode to be switched on before you can add a connector. Where that setting lives, and which plans include it, varies and changes — see OpenAI’s guide to Developer mode and MCP apps for current details.
For the technical folks: Claude Code, Cursor, Hermes & other MCP clients
Any MCP client that speaks Streamable HTTP can connect — Claude Code, Cursor, Codex, Hermes, or your own agent. Add this to your client’s MCP configuration:
{
"mcpServers": {
"adbadger": {
"url": "https://ai.adbadger.com/mcp/ads"
}
}
}
Authentication is OAuth with Dynamic Client Registration; your client will open the Ad Badger login on first use.
Verify it worked — ask your AI: “What Amazon Ads accounts do I have connected in Ad Badger?” If it lists your accounts, you’re live.
Want it to make changes too?
Every connection starts read-only. To let Claude or ChatGPT change bids, placement modifiers and budgets, and pause or enable campaigns, keywords, targets and product ads, re-authorize the Ad Badger connection from your AI app and allow changes when you sign in. It still asks before every change, and you can go back to read-only the same way.
Security
Read-Only Until You Say Otherwise
Read-Only by Default
A new connection cannot alter bids, change budgets or pause anything. Allow changes and it asks before every one, logs each one in your change history as “Badger MCP (AI)”, and still cannot delete anything.
Zero Credential Sharing
All access uses secure OAuth tokens. Your Amazon Seller Central password and AWS keys are never requested or stored by third parties.
Built on 9 Years of Amazon PPC Leadership
Ad Badger has pioneered Amazon PPC automation since 2017. We build tools to protect seller profitability, not inflate Amazon ad revenues.
Ad Badger is a verified partner in the official Amazon Ads Partner Directory.
Compare
Ad Badger vs. Amazon’s Own Amazon Ads MCP Server
Amazon Ads runs its own MCP server too. The two suit different people, so here is how to choose.
Choose Ad Badger for answers without the developer setup
- Connect with your Ad Badger account. No Amazon developer account, AWS credentials or API keys.
- Read-only until you allow changes, and it asks before every change.
- Built for seller questions, including n-gram wasted spend, placement comparisons, Search Query Performance, Total ACOS and organic rank.
Choose Amazon’s server if you are building on the Ads API
- Built and maintained by Amazon Ads.
- Setup starts with a Login with Amazon application and an Amazon developer account with Ads API access.
- A wider toolset, including campaign creation and reporting, with separate server URLs by region.
Details of Amazon’s server are from the Amazon Ads API documentation.
FAQ
Frequently Asked Questions
What is MCP?
MCP (Model Context Protocol) is an open standard that lets AI assistants like Claude and ChatGPT securely connect to outside tools and data. Ad Badger runs an Amazon Ads MCP server so your AI can query your live Amazon Ads data directly — no exports, no uploads, no copy-pasting CSVs.
Do I need an Amazon Ads API developer account or AWS credentials?
No. Ad Badger handles all developer access and API hosting. You only need your regular Ad Badger customer account with your Amazon store connected.
Does this cost extra?
No — the MCP connection works on every Ad Badger plan, free trials included.
Can Claude or ChatGPT make live changes or spend money?
Only if you allow it. Connections are read-only by default. If you re-authorize with changes allowed, Claude or ChatGPT can change bids, placement adjustments and daily budgets, and pause or enable campaigns, keywords, product targets and product ads (Sponsored Products only for now). It asks before every change, and every change is logged in your Ad Badger change history as “Badger MCP (AI)”. It cannot delete anything or create campaigns.
Why does data for the last 48 hours look incomplete?
Amazon Ads reporting naturally has an attribution lag of 24–48 hours. Our MCP connection transparently reflects this standard Amazon reporting window.
What if I have multiple brands or international marketplaces?
Ad Badger MCP natively supports all your connected accounts. You can simply ask: “Compare my US and UK account spend for last month” and receive side-by-side breakdowns in their respective currencies.
Is my data used to train AI models?
Ad Badger never uses your advertising data to train AI models. What Claude or ChatGPT do with your conversations is governed by your own AI account’s data settings — both offer training opt-outs.
How do I disconnect?
Remove the Ad Badger connector in your AI app at any time — in Claude under Settings ➞ Customize ➞ Connectors, or in ChatGPT’s connector settings. Disconnecting instantly stops all access; nothing keeps running in the background.
Tracking the Buy Box too? There’s an MCP for that: Buy Box Checker MCP — from the same team.
Talk to Your Amazon Ads in Plain English Today
Connect your Ad Badger account to Claude or ChatGPT in less than 60 seconds. Included on every plan, trials too. The Badger is already in the burrow, paws on your numbers.
Instant OAuth login • Read-Only by Default • Zero AWS Configuration • Maximum Badger
Claude is a trademark of Anthropic (claude.ai). ChatGPT is a trademark of OpenAI (chatgpt.com). Hermes Agent is a product of Nous Research (hermes-agent.nousresearch.com). All marks are property of their respective owners; Ad Badger is not affiliated with or endorsed by them.
Last updated: November 2, 2026