For AI users & developers
Connect Pinksheep to Cursor, Claude, and the AI tools you already use.
Quick answer
Pinksheep MCP lets you use your business tools inside MCP-compatible AI clients through agent-level endpoints. You connect the tools once in Pinksheep, expose the right agent over MCP, and keep approval controls, spend limits, and full history in front of important actions.
Pinksheep lets you use your business tools inside Cursor, Claude Desktop, and other MCP-compatible AI tools, with approval controls and full history on every action.
Use Pinksheep agents directly inside Cursor, Claude Desktop, and other AI tools. Your business tools stay connected, with approval controls, full visibility, and spend limits built in.
By pinksheep Editorial Team | Updated 25 March 2026
What is an MCP server?
Definition
Think of it as a way to plug your business tools into any AI you use. Instead of your AI only knowing what's in its training data, it can actually read from your HubSpot, write to your sheets, and take real actions through Pinksheep, where you keep oversight.
Why it matters
MCP (Model Context Protocol) is an open standard. Cursor, Claude, Windsurf, and other AI tools support it natively, which means one Pinksheep connection works across all of them.
How Pinksheep uses MCP
Pinksheep sits between your AI tools and your business data. Any AI client can call your tools, with Pinksheep's approval controls and full history in front of every action.
MCP servers by tool
Full directory →Jira MCP Server
Project MgmtRead issues, search sprints, propose ticket updates from Cursor or Claude.
Slack MCP Server
CommsRead channels, search messages, propose Slack posts with approval gates.
HubSpot MCP Server
CRMQuery contacts, read pipelines, approve CRM updates from your AI tool.
Salesforce MCP Server
CRMQuery Salesforce orgs and propose record updates with human sign-off.
Notion MCP Server
Project MgmtRead databases, search pages, propose property updates from Cursor.
Cursor MCP Server
DevConnect your business tools to Cursor via Pinksheep.
Common questions
What is an MCP server?
An MCP server (Model Context Protocol server) is a standardised interface that lets AI tools like Cursor and Claude connect to external tools and data sources. Pinksheep publishes MCP servers for your business tools: Jira, Slack, HubSpot, Salesforce, Notion, and hundreds more.
What is the difference between an MCP server and an API?
An API connects one specific thing to another. An MCP server is a standard that any AI tool (Cursor, Claude, Windsurf) can use to access your tools without requiring a custom integration for each one. One Pinksheep MCP connection works across all MCP-compatible AI clients.
What AI tools does Pinksheep connect to?
Cursor, Claude Desktop (Anthropic), Windsurf, and any other MCP-compatible AI tool.
Does my AI need my approval to make changes?
Pinksheep sits between your AI tool and your business data. You decide which actions need your sign-off before they run, and every important action is logged in the audit trail.
How do I connect Pinksheep to Cursor or Claude?
Connect your tools in Pinksheep, copy the MCP server URL, and paste it into your Cursor or Claude Desktop settings. It takes a few minutes.
Which MCP servers does Pinksheep provide?
Pinksheep provides MCP servers for Jira, Slack, HubSpot, Salesforce, Notion, GitHub, PostgreSQL, Google Sheets, Linear, Asana, Zendesk, and 500+ more via OAuth integrations. Every server includes a full audit trail and approval controls on important write actions.
What teams build with Pinksheep MCP
AI coding tools
Use your project management tools (Jira, GitHub, Linear) directly inside Cursor, with approval controls on any changes your AI makes.
CRM & sales tools
Let Claude or Cursor query and update your CRM through Pinksheep. Every field change is logged and approved before it goes through.
Finance & accounting tools
Give your AI tools read-access to financial data for analysis. Any changes require sign-off before anything updates.
Next step
Go from MCP overview to the specific deployment path
The overview page should hand off to the detail pages, integration surfaces, and comparison content that explain how MCP works in practice.
Browse the full MCP directory
Move from the overview page into individual tool and client pages for HubSpot, Slack, Jira, Cursor, and more.
IntegrationSee the integration layer
Understand the canonical tool integrations behind the MCP surfaces and agent workflows.
GuideRead the operator guide
See how teams deploy MCP access with governance, approvals, and a clear operating model.
CompareCompare MCP-ready options
See how pinksheep differs from direct agent frameworks when you need approvals and auditability.
Editorial and trust
MCP guidance is tied to real product setup paths
This MCP overview is published by the pinksheep Editorial Team and reviewed against current supported clients, product behaviour, and founder operating context so the setup claims are attributable and reviewable.
Published by
pinksheep Editorial Team
Product pages, guides, comparisons, and integration explainers are maintained as part of the pinksheep website editorial surface.
See the editorial teamReviewed against
Nick Hugh
Founder review anchors the product claims to real operating experience across CRM, systems, and software delivery work.
Review founder contextOperated by
Marshall Tech Group Pty Ltd
Sydney, Australia. Support: hello@pinksheep.ai. Legal and policy pages are published on the same site for verification.
Last reviewed 25 March 2026
Connect your tools to the AI you already use.
Set up Pinksheep. Connect your tools. Use them inside Cursor, Claude, and more.