How to Automate Pinterest Ads with Claude AI and MCP — E-Commerce Workflows, Shopping Ads, and Real Results

Pinterest is not a social network. It's a search engine for things people are planning to buy.

When someone opens Pinterest and searches "minimalist home office desk" or saves a pin for "wedding guest outfit ideas," they are not passively scrolling. They are in active planning mode — building wishlists, comparing options, and moving toward a purchase decision. That's why Pinterest users are 7x more likely to say Pinterest is the most influential platform in their purchase journey, compared to Facebook or Instagram.

For e-commerce advertisers, that purchase intent is extraordinarily valuable. The problem is that Pinterest Ads are genuinely underutilized — not because the platform doesn't work, but because most teams don't know how to read Pinterest's unique performance signals. Save rate. Closeup rate. Outbound click rate. These metrics tell a completely different story than CTR and CPM, and most advertisers either ignore them or never pull them out of Ads Manager at all.

The Pinterest Ads MCP server changes that. It gives Claude a live, authenticated connection to your Pinterest for Business account — so instead of manually navigating between campaign, ad group, and Pin-level views, you ask Claude which products are converting, which Pins are saving but not clicking, and where your catalog spend is leaking.

---

What Is the Pinterest Ads MCP Server?

The Pinterest Ads MCP server implements the Model Context Protocol — the open standard that lets AI models connect to external tools and data sources with live, authenticated access.

It bridges Claude directly to the Pinterest Marketing API. Ask about campaign performance — Claude queries your live Ads account. Ask which Pins have strong save rates but weak outbound clicks — Claude pulls the full Pin-level breakdown. Ask which products in your Shopping catalog are underperforming — Claude cross-references catalog data with conversion metrics and surfaces the SKUs to cut or scale.

What makes Pinterest's data uniquely valuable — and uniquely underused — is the intent layering. Pinterest gives you three distinct engagement signals on every ad:

Most advertisers only look at CTR and ROAS. Claude with the Pinterest MCP server lets you interrogate all three intent layers simultaneously — across every Pin, ad group, and campaign — in a single conversation.

---

Three Server Options Compared

1. Adzviser — Managed Remote MCP (Easiest, Multi-Platform)

Adzviser provides the fastest path to Pinterest Ads data in Claude. Connect your Pinterest account via OAuth, add the Adzviser MCP server URL to Claude Desktop, and start querying. The standout advantage: Adzviser supports Pinterest, Google Ads, and Meta Ads in a single MCP session — so Claude can analyze cross-platform performance without switching contexts.

Best for: E-commerce teams and agencies running Pinterest alongside Google and Meta who want unified cross-platform analysis without any technical setup.

2. CData Pinterest MCP Server — Full CRUD, Enterprise-Grade

CData offers two tiers. The open-source read-only version on GitHub connects to Pinterest data via CData JDBC drivers — good for analytics. The paid full version adds complete create, update, delete, and action capabilities — campaign management, ad group edits, and direct write operations against your live account. Supports all major Pinterest Ads entities: Campaigns, AdGroups, Ads, AdAccounts, AdAnalytics, AdCampaignAnalytics, and AdGroupAnalytics.

Best for: Enterprise teams and developers who need full read-write access and SQL-level control over Pinterest Ads data.

3. Insightful Pipe — Analytics-Focused, Automated Reports

Insightful Pipe is purpose-built for analytics workflows: pull Pinterest performance metrics, generate automated optimization reports, and track Pin performance, audience insights, and shopping trends. Connects multiple ad platforms for cross-channel reporting in a single MCP session.

Best for: Agencies and media buyers who need automated Pinterest performance reporting without managing infrastructure.

---

Tool Surface — What Claude Can Access

Pinterest Ads MCP servers expose tools across the full campaign hierarchy — accounts, campaigns, ad groups, Pins, and reporting — plus Pinterest-specific data like catalog analytics and audience insights:

Account & Campaign Management

Tool	What It Does
get_ad_accounts	List all Pinterest ad accounts your credentials can access
get_campaigns	Retrieve campaigns with objective, status, and lifetime/daily budget
create_campaign	Launch a new campaign with objective (awareness, consideration, conversions, catalog sales)
update_campaign	Modify budget, status, name, or campaign objective

Ad Group Operations

Tool	What It Does
get_ad_groups	Retrieve ad groups with targeting, bid strategy, and spend limits
create_ad_group	Create an ad group with audience, keyword, and interest targeting
update_ad_group	Adjust bid, targeting, schedule, or status
get_targeting_options	Browse available interest categories, demographics, and keyword suggestions

Pin & Creative Management

Tool	What It Does
get_ads	List Promoted Pins with creative details, format, and delivery status
create_ad	Promote a Pin as an ad within an ad group
update_ad	Modify ad status or creative association
get_ad_details	Full Pin creative specs, destination URL, and delivery settings

Performance Analytics — Pinterest-Specific

Tool	What It Does
get_campaign_analytics	Campaign-level performance: impressions, saves, closeups, outbound clicks, spend, ROAS, CPA, conversions
get_ad_group_analytics	Ad group performance with the same metric set — compare audience segments and targeting approaches
get_ad_analytics	Pin-level performance breakdown — identify top saves, best outbound click rates, and highest-ROAS creative
get_audience_insights	Demographic performance breakdown by age, gender, location, and device

Shopping & Catalog

Tool	What It Does
get_catalogs	List product catalogs connected to your Pinterest Business account
get_catalog_products	Browse catalog items with product-level performance data
get_product_group_analytics	Performance breakdown by product group — identify which SKU clusters are converting and which are wasting spend

---

Setup: Two Paths to Connection

Path 1 — Adzviser (5 Minutes, Managed)

1. Sign up at adzviser.com
2. Connect your Pinterest Ads account via OAuth
3. Copy your Adzviser MCP server URL
4. Add to Claude Desktop config at ~/.config/claude/claude_desktop_config.json

{
  "mcpServers": {
    "pinterest-ads": {
      "url": "https://mcp.adzviser.com/pinterest?api_key=YOUR_ADZVISER_KEY"
    }
  }
}

Path 2 — CData Pinterest MCP Server (Self-Hosted)

# Clone and install the CData read-only server
git clone https://github.com/CDataSoftware/pinterest-mcp-server-by-cdata
cd pinterest-mcp-server-by-cdata
npm install

Then add to Claude Desktop config:

{
  "mcpServers": {
    "pinterest-ads": {
      "command": "node",
      "args": ["/path/to/pinterest-mcp-server-by-cdata/index.js"],
      "env": {
        "PINTEREST_ACCESS_TOKEN": "your_access_token_here",
        "PINTEREST_AD_ACCOUNT_ID": "your_ad_account_id"
      }
    }
  }
}

To get your Pinterest access token: go to developers.pinterest.com, create an app, and request Ads Read (or Ads Write for full CRUD) scopes.

---

Workflow 1 — Morning Performance Audit (Reading Pinterest's Unique Signals)

You: "Give me a 7-day performance summary across all active Pinterest
campaigns. For each ad group, show ROAS, save rate, outbound click rate,
and spend. Flag anything with ROAS below 2.0 or save rate above 4%
but outbound click rate below 0.5%."

Claude pulls data via get_campaign_analytics and get_ad_group_analytics:

"Here's your 7-day Pinterest Ads summary across 4 active campaigns:

🔴 NEEDS ATTENTION:
- 'Home Decor — Consideration' — Save rate: 6.2%, Outbound CTR: 0.3%
  High saves, very low outbound clicks. Users love the visuals but the
  destination URL or product price may be creating friction.
  Recommendation: A/B test the landing page and review product pricing.

- 'Fashion — Conversions' — ROAS: 1.6, Spend: $280/day
  Below target ROAS. Ad group targeting broad interest — insufficient
  purchase intent signals. Recommend shifting to keyword targeting
  around specific product searches.

✅ PERFORMING WELL:
- 'Shopping — Home Goods Catalog' — ROAS: 5.1, Outbound CTR: 1.8%
  Strong performer. Catalog shopping ads converting well on home goods.
  Budget constrained — candidate for 30% spend increase.

- 'Beauty — Retargeting' — ROAS: 6.3, Save rate: 2.1%
  Highest ROAS in account. Small audience — monitor frequency."

Workflow 2 — Save Rate vs. Outbound Click Analysis (Pinterest's Hidden Intelligence)

Pinterest's save rate is the most underutilized metric in paid social advertising. A high save rate means users are actively bookmarking your content into their planning boards — a level of purchase intent that no other platform can measure. But save rate alone doesn't mean you're capturing that intent. The save vs. outbound click gap tells you exactly where the funnel is breaking.

Three patterns tell different stories:

You: "For all active Pins in the 'Home Decor' campaign, show save rate
and outbound click rate side by side for the last 14 days. Classify
each Pin as high-save-high-click, high-save-low-click, or low-save-low-click.
Tell me which landing pages the high-save-low-click Pins are pointing to."

Claude calls get_ad_analytics and get_ad_details:

"Pin performance classification — 'Home Decor' campaign (14 days):

HIGH SAVE + HIGH OUTBOUND CLICK ✅ (scale):
- Pin 3847 'Minimalist desk setup' — Save: 5.8%, Outbound: 1.9% — $48 CPA
- Pin 3851 'Neutral bedroom palette' — Save: 4.2%, Outbound: 1.6% — $52 CPA

HIGH SAVE + LOW OUTBOUND CLICK ⚠️ (fix landing page):
- Pin 3849 'Rattan shelving unit' — Save: 7.1%, Outbound: 0.2%
  Landing page: yoursite.com/collections/shelving
  The visual is highly engaging (7% save rate is exceptional).
  Issue is likely post-click: page load time, product availability,
  or price mismatch. Recommend reviewing landing page experience.

LOW SAVE + LOW OUTBOUND CLICK ❌ (replace):
- Pin 3852 'Living room inspo' — Save: 0.9%, Outbound: 0.3%
  Neither engagement signal is strong. Creative is not resonating.
  Recommend pausing and replacing with a product-specific visual."

Workflow 3 — Shopping Ads Catalog Optimization

Pinterest Shopping Ads are the platform's highest-ROAS format for e-commerce. Unlike social feed ads where creative quality drives performance, Shopping Ads serve products dynamically from your catalog — so the levers are catalog coverage, product group structure, and bid strategy by category.

Most advertisers running Shopping Ads are flying blind: they know total ROAS but not which product groups, price tiers, or categories are driving it — and which are dragging it down. Claude with the Pinterest MCP server makes catalog analytics conversational.

You: "Analyze my Pinterest Shopping campaign performance by product group
for the last 30 days. Which product groups are below our target ROAS
of 3.0? Which have strong ROAS but low impressions — suggesting
they're not getting enough budget?"

Claude calls get_product_group_analytics:

"Shopping catalog analysis — last 30 days:

SCALE CANDIDATES (strong ROAS, budget constrained):
- Product Group: Rugs & Floor Coverings — ROAS: 6.2, Impressions: 8,400
  Low impression share relative to performance. Increase bid floor
  or daily budget cap for this group.
- Product Group: Lighting — ROAS: 4.8, Impressions: 5,200
  Same pattern. High-intent category underserved by current budget.

BELOW TARGET ROAS (reduce or exclude):
- Product Group: Seasonal Decor — ROAS: 1.4, Spend: $380
  Off-season products likely driving poor conversion. Recommend
  excluding this group until seasonal relevance returns.
- Product Group: Clearance Items — ROAS: 1.1, Spend: $210
  Clearance pricing may not align with Pinterest's high-AOV audience.
  Recommend testing a price floor exclusion filter in the product feed.

MISSING FROM CATALOG: 0 products excluded due to feed errors ✅"

Workflow 4 — Keyword Targeting Intelligence

Pinterest is the only major ad platform where keyword targeting works like a search engine. Users actively type queries — "small bathroom renovation ideas," "capsule wardrobe fall 2026," "birthday cake recipes for adults" — and their search intent is extremely high because they're planning, not scrolling.

This means Pinterest keyword targeting gives you something Meta and TikTok cannot: the ability to reach users at the exact moment of category intent. But it also means keyword selection matters as much as it does in Google Ads — and most Pinterest advertisers are targeting too broadly.

You: "For our 'Skincare' campaign, I want to understand how our current
keyword targeting is performing. Which keywords have the best outbound
click rate? Which are generating saves but no clicks? And are there
high-intent keyword categories we're missing?"

Claude analyzes via get_ad_group_analytics with keyword breakdown:

"Keyword performance — 'Skincare' campaign:

TOP PERFORMERS (highest outbound CTR):
- 'vitamin c serum for face' — Outbound CTR: 2.4%, CPA: $18
- 'best moisturizer for dry skin' — Outbound CTR: 2.1%, CPA: $21
- 'retinol before and after' — Outbound CTR: 1.8%, CPA: $24

HIGH SAVE, LOW CLICK (landing page opportunity):
- 'skincare routine steps' — Save: 8.2%, Outbound CTR: 0.3%
  Very high planning intent — users are researching routines.
  Consider a dedicated routine-building landing page that educates
  before asking for the conversion.

GAPS — High-volume keywords not currently targeted:
- 'glass skin routine'
- 'korean skincare routine for beginners'
- 'niacinamide serum benefits'
  These are top search trends in the beauty category with strong
  purchase signal. Recommend testing in a new ad group."

Workflow 5 — Budget Reallocation by ROAS Across Campaigns

An apparel brand was running $3,000/day across five Pinterest campaigns — brand awareness, consideration, conversions, shopping, and retargeting — at roughly equal budget. The data after 30 days showed a familiar pattern: two campaigns were generating nearly all revenue, while two were consuming budget with no meaningful conversion contribution.

You: "Analyze the last 30 days of Pinterest performance across all
campaigns. What is the ROAS and spend for each? Which are above our
target ROAS of 3.5 and which are below? Recommend a budget reallocation."

Claude pulls via get_campaign_analytics:

"30-day campaign performance — full account:

Campaign              Spend       ROAS    vs. Target
Brand Awareness       $900/day    N/A     Awareness — no ROAS target
Consideration         $600/day    1.2x    ❌ Below target
Conversions           $600/day    4.6x    ✅ Above target
Shopping Catalog      $600/day    5.8x    ✅ Above target
Retargeting           $300/day    7.2x    ✅ Above target

Recommended reallocation ($3,000/day total):
- Brand Awareness: $700/day (slight reduction — audience saturating)
- Consideration: $300/day (reduce — low efficiency, rebuild with keyword targeting)
- Conversions: $800/day (increase — strong performer, not spend-constrained)
- Shopping Catalog: $900/day (increase — highest-volume opportunity)
- Retargeting: $300/day (maintain — small audience, don't over-serve)

Projected impact: +28–35% total conversions at same spend level,
based on current ROAS differential between reallocation sources."

The brand implemented the reallocation. Over the following 30 days: total Pinterest-attributed revenue increased 31% with identical daily spend. The Shopping Catalog budget increase alone added $14,000 in monthly revenue.

---

Benchmarks: What Pinterest Advertisers Are Reporting

---

What Claude Cannot Do — Be Realistic

The right mental model: Claude is your Pinterest data analyst working at AI speed. It surfaces what's in your account, what the patterns mean, and what to do next. The creative judgment, catalog strategy, and final approval on any spend change remain with you.

---

Combining Pinterest Ads MCP with Other Servers

Pinterest's highest value for most e-commerce brands comes when it's analyzed alongside other channels and data sources:

---

Getting Started Today

Frequently Asked Questions

Related Guides

• How to Automate Meta Ads with Claude AI and MCP
• How to Automate Google Ads with Claude AI and MCP
• How to Automate Amazon Ads with Claude AI and MCP
• How to Automate LinkedIn Ads with Claude AI and MCP
• How to Automate TikTok Ads with Claude AI and MCP

---

Pinterest Ads MCP integrations are unofficial tools not affiliated with or endorsed by Pinterest, Inc. Always review Pinterest's Marketing API terms of service for your use case.

---

Resources:

• Adzviser Pinterest Ads MCP — Managed remote MCP, multi-platform support, no API credentials required
• CDataSoftware/pinterest-mcp-server-by-cdata — Read-only, free, open source
• CData Pinterest MCP Server (Full CRUD) — Full read-write access, enterprise-grade
• Insightful Pipe Pinterest MCP — Analytics-focused, automated reports
• Pinterest Marketing API — Official developer documentation
• MCP Playground Online — Test MCP servers in your browser