MCP Server

Salesforce + Pardot MCP Server

Remote MCP server combining Salesforce CRM and Pardot (Marketing Cloud Account Engagement) data in 22 tools — pipeline reports, prospect scoring, email analytics, and visitor activity via OAuth 2.1.

8

Installation

git clone https://github.com/shelfio/salesforce-pardot-mcp && cd salesforce-pardot-mcp && pip install -r requirements.txt && python server.py
Platform
Claude Code, Claude Desktop, Cursor
Pricing
free
Author
shelfio
Last Tested
2026-04-01

About this Tool

The only MCP server that combines Salesforce CRM pipeline data with Pardot (now called Salesforce Marketing Cloud Account Engagement) marketing automation in a single connection. Provides 22 tools across both platforms: on the Salesforce side — SOQL query execution, lead/contact management, pipeline reporting, task and event history, and lead creation; on the Pardot side — prospect lookup with score/campaign filtering, campaign and list enumeration, form tracking, email analytics, visitor activity with enriched event labels (form submits, email opens/clicks/bounces, page views), lifecycle stage history, and list membership management. Authentication uses OAuth 2.1 with PKCE — Claude Desktop opens a standard OAuth popup to your Salesforce org, so end-users connect with a click rather than managing API tokens. Deployable on Railway or run locally via pip. Actively maintained with security hardening commits in April 2026.

Use Case Examples

  • Pull all Pardot prospects with lead score above 75 and cross-reference against open Salesforce opportunities
  • Report on email campaign engagement (opens, clicks, bounces) for a specific Pardot campaign
  • Query Salesforce pipeline by stage and forecast category to brief a marketing team on deal velocity
  • Identify prospects who submitted a form but have no associated Salesforce contact — potential data quality issue
  • Review lifecycle stage progression history for top-scoring prospects to understand conversion patterns

MarketCore Editorial Review

MarketCore Editorial Review

The enterprise B2B marketing ops gap in the catalog is now filled. Salesforce + Pardot is the dominant stack at mid-market and enterprise companies — marketing ops teams in these organizations have been waiting for an AI-native way to query Pardot prospect data alongside Salesforce pipeline data. This server delivers that. The OAuth 2.1 PKCE authentication is the right call for enterprise environments: no API keys to rotate, no service accounts to manage, and any team member can self-serve their connection. The read-oriented default is a significant trust advantage for enterprise IT. Actively maintained by Shelf.io, a developer studio with a track record of open-source contributions. The April 2026 security commits signal this is production-intent software, not a proof-of-concept.

Best For

  • Marketing-to-sales handoff analysis: find Pardot prospects with high scores that have not been converted to Salesforce opportunities
  • Campaign attribution: trace which Pardot email campaigns are generating the most Salesforce pipeline
  • Lead scoring audits: compare Pardot score distributions across campaigns to calibrate scoring models
  • Prospect engagement timelines: review full visitor activity history for key accounts before sales calls

Code Quality

FastMCP-based Python server with OAuth 2.1 PKCE, Railway deployment support, 22 well-defined tools with enriched event label mapping for Pardot visitor activities. Recent security hardening commits in March-April 2026 indicate active production-readiness work. Shelf.io is a credible author with other maintained OSS projects.

Limitations

Requires Salesforce Connected App setup (not trivial for non-admins). Pardot integration requires both Pardot and Salesforce org access. Railway deployment setup adds hosting complexity. No npm package — must clone and run locally or deploy to Railway. Zero stars currently (very new) but the code quality and active maintenance suggest this will gain traction.

Reviewed 2026-04-01

Score Breakdown

8

Overall Score

Based on 5 dimensions

Installation
6/10
Documentation
7/10
Relevance
9/10
Quality
8/10
Maintenance
8/10

Get weekly updates on new marketing AI tools

New tools reviewed every week. No spam.