Forem: CuratedMCP

Windsor.ai MCP Server: Query 325+ Marketing & Sales Data Sources in Plain English

curatedmcp — Sat, 16 May 2026 10:28:09 +0000

Windsor.ai MCP Server: Query 325+ Marketing & Sales Data Sources in Plain English

Stop writing SQL queries or toggling between dashboards. Windsor.ai MCP connects Claude, Cursor, and other MCP-compatible AI agents directly to your marketing, sales, and customer data across 325+ platforms—Google Ads, Meta, TikTok, GA4, HubSpot, Salesforce, Shopify, and more. Ask natural-language questions and get structured answers in seconds.

What It Does

Windsor.ai MCP bridges your marketing and sales tech stack to your AI agent. Instead of exporting CSVs or piecing together reports manually, you can ask Claude questions like:

"What campaigns had the best ROAS last month?"
"Break down spend by channel over the last 90 days"
"Which product had the highest conversion rate this quarter?"

The server handles authentication via your Windsor.ai account and API key, then translates plain English into queries across your connected sources. Responses come back structured and actionable—no dashboards, no custom ETL, no SQL required.

This is especially useful for building AI-powered analytics workflows, automating report generation, or letting non-technical teammates query their own data through Claude.

How to Install

Add the Windsor.ai MCP server to Claude Desktop (Pro+ plan), Cursor, or any MCP-compatible client:

https://mcp.windsor.ai

For Claude Desktop, add this to your claude_desktop_config.json:

{
  "mcpServers": {
    "windsor.ai-mcp-server": {
      "command": "https://mcp.windsor.ai"
    }
  }
}

You'll need a Windsor.ai account and API key. The server is free during beta.

Real-World Use Cases

Automated Weekly Reports: Ask Claude to summarize campaign performance across all channels and generate a plain-text report you can paste into Slack or email.
Quick Ad Audit: Query spend, impressions, and conversion data across Google Ads, Meta, and TikTok in one shot—useful for spotting budget leaks or underperforming creatives.
Attribution Analysis: Compare ROAS, CPA, or customer lifetime value across acquisition channels without switching tools or waiting for a data analyst.

Full install guides for Claude Desktop, Cursor, Windsurf, and more at CuratedMCP.

Perspective AI: Replace Static Forms with Conversational Data Collection

curatedmcp — Fri, 15 May 2026 10:14:32 +0000

Perspective AI: Replace Static Forms with Conversational Data Collection

Static forms kill engagement. They're rigid, impersonal, and lose context before users even submit. Perspective AI is an MCP server that swaps forms for adaptive AI conversations—letting your AI agents conduct intelligent interviews that capture structured data while maintaining the full situational context.

Instead of forcing users through checkbox-and-dropdown workflows, Perspective AI creates dynamic conversations that understand intent, follow natural dialogue patterns, and automatically structure the information that actually matters. Think of it as deploying an AI concierge that knows when to dig deeper, when to summarize, and when to move to the next step.

What It Does

The Perspective AI MCP server unlocks two core capabilities in Claude, Cursor, or Windsurf:

Adaptive Data Collection — Your AI agent conducts natural conversations that feel like talking to a knowledgeable person, not filling out a form. The system intelligently structures responses and routes follow-up questions based on context, not rigid logic trees.

Automation Triggers — Once structured data is captured, you can immediately trigger downstream actions: create tickets, fire webhooks, qualify leads, or log research findings—without manual data transformation.

This is especially powerful for customer research, application intake, support triage, and survey workflows where context dropout and weak qualification are expensive problems.

How to Install

Run the install command:

npx mcp-remote https://getperspective.ai/mcp

Then add to your claude_desktop_config.json:

{
  "mcpServers": {
    "perspective-ai": {
      "command": "npx mcp-remote https://getperspective.ai/mcp"
    }
  }
}

Real-World Use Cases

Customer Onboarding Interviews — Have Claude conduct a multi-turn conversation to understand a prospect's needs, automatically extract intent signals and pain points, and create a pre-qualified lead record in your CRM.
Structured User Research — Deploy an adaptive interviewer in Cursor that converses naturally with users about their workflows, auto-captures key findings in JSON, and flags patterns for product decisions.
Support Intake Triage — Let your AI agent ask clarifying questions in natural language, gather full context about a support issue, and route to the right team with structured data—reducing back-and-forth.

Full install guides for Claude Desktop, Cursor, Windsurf, and more at CuratedMCP.

DataForSEO MCP Server: Real-time SEO data and keyword intelligence for AI agents

curatedmcp — Thu, 14 May 2026 10:14:34 +0000

DataForSEO MCP Server: Real-time SEO data and keyword intelligence for AI agents

DataForSEO MCP Server bridges the gap between AI assistants and enterprise SEO intelligence. If you're building Claude, Cursor, or Windsurf integrations that need to pull keyword metrics, SERP rankings, backlink data, or on-page performance insights, this is your direct pipe to DataForSEO's APIs without writing custom integrations.

What It Does

The server implements a standardized Model Context Protocol interface to DataForSEO's comprehensive suite of SEO and marketing data APIs. Your AI agent gains access to:

Real-time SERP data from Google, Bing, and Yahoo for rank tracking and competitor analysis
Keyword research metrics including search volume, CPC, and trend data
On-page crawling with customizable parameters to audit website SEO performance
Backlink intelligence covering referring domains, anchor text, and link quality scores
Domain analytics for traffic patterns, technology stacks, and Whois details
AI Optimization insights for keyword discovery and LLM benchmarking
Business data and content analysis for brand monitoring and sentiment tracking

This means your Claude instance can autonomously fetch live keyword rankings, analyze competitor backlinks, or audit a website's technical SEO—all through natural conversation. No manual API calls, no authentication headaches.

How to Install

npx dataforseo-mcp-server

Add to your Claude Desktop configuration:

{
  "mcpServers": {
    "dataforseo-mcp-server": {
      "command": "npx dataforseo-mcp-server",
      "env": {
        "DATAFORSEO_USERNAME": "your_username",
        "DATAFORSEO_PASSWORD": "your_password"
      }
    }
  }
}

Grab your credentials from DataForSEO.

Real-World Use Cases

Competitive keyword analysis: Ask Claude to pull top-ranking keywords for a competitor domain, get search volumes and CPC data in seconds, then receive structured recommendations for content gaps.
Technical SEO audits: Have Claude crawl your staging environment via the On-Page API, surface critical on-page issues (missing meta tags, broken internal links, schema problems), and generate a prioritized fix list.
Backlink opportunity research: Claude queries the Backlinks API for high-authority sites linking to competitors but not to you, then surfaces outreach targets with anchor text distribution and domain metrics built in.

Full install guides for Claude Desktop, Cursor, Windsurf, and more at CuratedMCP.

HubSpot MCP: Give Claude Direct Access to Your CRM and Revenue Data

curatedmcp — Wed, 13 May 2026 10:14:33 +0000

HubSpot MCP: Give Claude Direct Access to Your CRM and Revenue Data

HubSpot MCP is the official Model Context Protocol server from HubSpot. It connects AI agents like Claude, Cursor, and Windsurf directly to your HubSpot account, letting them read and write CRM data, manage workflows, and pull analytics—all without leaving your IDE or chat interface.

What It Does

This server unlocks a new layer of automation for revenue teams. Instead of switching between HubSpot and your code editor, you can ask Claude to create contacts, update deals, manage support tickets, and generate follow-up tasks in context. The server bridges your CRM to AI reasoning, enabling agents to:

Create and update contacts, companies, and deals with full property support
Manage customer support by creating, updating, and triaging tickets
Query campaign performance and pull email analytics
Access sales pipeline data and generate revenue reports
Auto-generate follow-ups and email drafts based on customer context
Maintain lists and custom properties programmatically

This is useful if you're building internal tools, automating lead qualification, or letting AI handle routine CRM hygiene while you focus on strategy.

How to Install

Install via npm:

npx -y @hubspot/mcp

Then add it to your Claude Desktop config (or Cursor/Windsurf):

{
  "mcpServers": {
    "hubspot-mcp": {
      "command": "npx -y @hubspot/mcp",
      "env": {
        "HUBSPOT_API_KEY": "your-api-key-here"
      }
    }
  }
}

Generate a private app API key in HubSpot Settings → Integrations → Private Apps, then paste it as an environment variable.

Real-World Use Cases

Lead enrichment automation: Ask Claude to fetch a contact's recent activity, update their company info, and move their deal stage based on email engagement data—all in one request.
Support ticket triage: Have your AI agent pull open tickets, categorize them by priority, assign them to the right team, and draft responses based on conversation history.
Weekly pipeline reports: Generate a revenue snapshot by querying all deals in specific pipeline stages, with win/loss analysis and next-step recommendations formatted for your team.

Full install guides for Claude Desktop, Cursor, Windsurf, and more at CuratedMCP.

Databricks MCP: Give Claude Direct Access to Your Lakehouse

curatedmcp — Tue, 12 May 2026 10:14:33 +0000

Databricks MCP: Give Claude Direct Access to Your Lakehouse

If you're running analytics or data engineering workflows on Databricks, you've probably wished your AI tools could actually do something with all that data instead of just talking about it. Databricks MCP closes that gap.

This is the official integration that lets Claude, Cursor, and other AI agents execute notebooks, query Delta tables, manage clusters, and inspect your data lineage—all from within your chat. Think of it as giving your AI a direct line to your lakehouse instead of just read-only documentation.

What It Does

Databricks MCP unlocks three core capabilities:

Data access. Query Delta Lake tables and Unity Catalog schemas directly. Your AI can write SQL, inspect table metadata, and understand your data structure without context-switching to the Databricks UI.

Automation. Execute notebooks and run SQL analytics on demand. Instead of manually kicking off jobs, you can ask Claude to run a transformation, check the results, and iterate—all in conversation.

Compute & ML ops. Manage cluster state, access MLflow experiments and model registries, and inspect job histories. Your AI becomes a useful teammate for routine ops tasks.

The server handles authentication via Databricks credentials and integrates seamlessly into existing MCP-compatible tools.

How to Install

Install via npm:

npx -y @databricks/mcp

Then add to your Claude Desktop config (typically ~/.config/Claude/claude_desktop_config.json on Linux/Mac):

{
  "mcpServers": {
    "databricks-mcp": {
      "command": "npx -y @databricks/mcp",
      "env": {
        "DATABRICKS_HOST": "your-workspace-url",
        "DATABRICKS_TOKEN": "your-pat-token"
      }
    }
  }
}

Store credentials securely—never commit tokens to version control. Refer to the official docs for Cursor and Windsurf setup.

Real-World Use Cases

Debugging pipeline failures. "Run this notebook and show me what broke in the transformation step"—get results in seconds instead of manually checking the job UI.
Ad-hoc analysis. "Query our customer table, count rows by region, and highlight anomalies"—your AI writes the SQL and surfaces insights without you touching the cluster.
Model iteration. "Check the latest MLflow run, compare metrics to the last three experiments, and suggest which one to promote to staging"—keep experiments and context in one place.

Full install guides for Claude Desktop, Cursor, Windsurf, and more at CuratedMCP.

MCP Ecosystem Week 20: Database Tooling Enters the Ring as Giants Dominate Discovery

curatedmcp — Mon, 11 May 2026 10:17:57 +0000

MCP Ecosystem Week 20: Database Tooling Enters the Ring as Giants Dominate Discovery

The MCP ecosystem continues its steady expansion with the addition of Local-YDB, an unofficial server that brings local YDB database management into the MCP fold. Meanwhile, the weekly trending list reveals a clear pattern: integration servers backed by major platforms—OpenAI, GitHub, Figma, and Anthropic—continue to capture developer attention at scale.

This Week in MCP

We added one new server this week:

Local-YDB unofficial mcp server (FREE): A community-built server for managing local YDB database instances. YDB, Yandex's distributed SQL database, joins the growing roster of database-specific MCP tools as developers seek to integrate data layer operations directly into their AI workflows.

The addition pushes the verified catalog to 70 servers, with 69 remaining free. No new playbooks were added this week, keeping the focus on server infrastructure expansion rather than orchestration patterns.

Trending Servers

GitHub Copilot MCP continues its dominant run, commanding 98K views this week. The server channels GitHub's code intelligence—completions, explanations, and review capabilities—into any MCP client, making it an obvious draw for developers working in Cursor and other AI IDEs. It's the bridge between GitHub's ecosystem and broader AI agent workflows.

OpenAI MCP lands at 87K views, offering official access to GPT-4o, DALL-E, Whisper, and Embeddings APIs through MCP. The steady demand signals that developers want to orchestrate multiple OpenAI capabilities within unified agent contexts rather than making isolated API calls.

Figma MCP rounds out the top three with 82K views, serving design engineers and teams building AI-assisted workflows around design systems. Direct access to design files, components, and design tokens from within an agent drastically reduces context-switching overhead—a compelling value proposition that the numbers clearly reflect.

The pattern here is worth noting: first-party integrations from well-known platforms capture the lion's share of attention. GitHub, OpenAI, Figma, and Anthropic collectively pull roughly 420K views across the top five slots. Developer friction and discoverability favor the names they already trust.

Quick Take

The 100% free-tier catalog (69 of 70 servers) masks an important reality: pricing structures around MCP are still forming. No commercial/premium servers have gained traction yet, suggesting the ecosystem remains in the "free tooling to build the foundation" phase. What's notable is how developers are using free servers—they're threading together specialized capabilities (database tools, design platforms, code hosts) into composite agents rather than adopting monolithic platforms. This suggests MCP's real value isn't in individual integrations, but in the ability to mix and match. The Local-YDB addition is a quiet marker: database teams, not just SaaS platforms, are beginning to see MCP as infrastructure worth supporting. Watch for more domain-specific tooling (observability, data, infrastructure) to follow as the pattern becomes clearer.

Browse all 70 verified servers at CuratedMCP.

PagerDuty MCP: Let AI Handle Your On-Call Chaos

curatedmcp — Sun, 10 May 2026 10:45:34 +0000

PagerDuty MCP: Let AI Handle Your On-Call Chaos

Operations teams spend too much time context-switching between incident dashboards, chat, and email. PagerDuty MCP is the official integration that gives Claude, Cursor, and other AI agents direct access to your incident response workflows—so they can create incidents, manage alerts, update on-call schedules, and keep responders aligned without human hand-offs.

What It Does

PagerDuty MCP bridges the gap between your AI agent and your incident management system. Instead of manually toggling between tools, you can now ask Claude to create an incident, check who's on-call, escalate a critical alert, or annotate a timeline—and the agent executes those commands directly in PagerDuty.

The server unlocks core capabilities: create and update incidents with custom urgency and priority levels, list active on-call users and rotation schedules, manage escalation policies and team assignments, query historical alerts and response timelines, annotate incidents with context or responder notes, and trigger incident actions programmatically from plain-language AI commands.

This is especially powerful for automation. When your monitoring system fires a critical alert, an AI agent can automatically open a PagerDuty incident, notify the right on-call engineer, and add diagnostic context—all in seconds. Or during an active incident, your agent can pull recent alert history, suggest escalation paths, and track response metadata without anyone leaving their terminal.

How to Install

npx -y @pagerduty/mcp

Add the server to your Claude Desktop config:

{
  "mcpServers": {
    "pagerduty-mcp": {
      "command": "npx -y @pagerduty/mcp",
      "env": {
        "PAGERDUTY_API_TOKEN": "your-api-token-here"
      }
    }
  }
}

You'll need a valid PagerDuty API token with appropriate scopes. Get yours from your PagerDuty account settings.

Real-World Use Cases

Incident triage at scale: Ask Claude to review your last 50 incidents, identify patterns (e.g., recurring database timeouts), and propose new escalation policies without manually digging through the UI.
Automated alert response: When a monitoring tool detects anomalies, an AI agent creates a PagerDuty incident, pulls the on-call roster, checks recent similar incidents, and sends a pre-populated summary to the responder.
On-call coordination: Query who's currently on-call across multiple services, check schedule conflicts, and let Claude suggest load-balancing changes based on response time trends.

Full install guides for Claude Desktop, Cursor, Windsurf, and more at CuratedMCP.

Splunk MCP: Query your observability stack directly from Claude

curatedmcp — Sat, 09 May 2026 10:45:37 +0000

Splunk MCP: Query your observability stack directly from Claude

Splunk MCP is the official Model Context Protocol server that bridges AI agents to your Splunk instance. Instead of context-switching between Claude and your monitoring platform, you can now ask your AI assistant to search logs, inspect alerts, correlate incidents, and surface root causes—all without leaving your chat.

What It Does

Splunk MCP gives Claude and other AI agents direct access to your observability and security data. You can run SPL (Splunk Processing Language) queries, retrieve event results, inspect logs across services, and access alert timelines. The server also lets you search saved dashboards, manage data inputs, and correlate security events with operational metrics.

The key unlock: your AI agent can now triage incidents autonomously. Feed it vague symptoms ("why did latency spike at 3pm?"), and it can query your Splunk instance, pull relevant logs and metrics, cross-reference alert history, and surface likely root causes. For SRE and SecOps teams drowning in data, this cuts investigation time significantly.

How to Install

Install via npm:

npx -y @splunk/mcp

Then add Splunk MCP to your Claude Desktop config:

{
  "mcpServers": {
    "splunk-mcp": {
      "command": "npx -y @splunk/mcp",
      "env": {
        "SPLUNK_URL": "https://your-splunk-instance.com",
        "SPLUNK_USERNAME": "your_username",
        "SPLUNK_PASSWORD": "your_api_token"
      }
    }
  }
}

Replace placeholders with your Splunk credentials and instance URL. Full setup guides for Cursor, Windsurf, and other clients available in the official docs.

Real-World Use Cases

Incident triage at speed: "Hey Claude, I got a PagerDuty alert for high error rates. Search Splunk logs from the last 30 minutes and tell me which service is failing." Claude queries your instance, correlates events, and suggests remediation steps.
Root cause analysis: "Why did our database queries slow down yesterday around 2pm?" Claude pulls metrics, compares baselines, and surfaces anomalies you'd normally miss.
Security event correlation: "Find all failed login attempts from the last 48 hours and correlate them with successful logins from unusual IPs." Claude runs complex multi-field searches without you writing SPL.

Full install guides for Claude Desktop, Cursor, Windsurf, and more at CuratedMCP.

ServiceNow MCP: Automate ITSM workflows without leaving your AI agent

curatedmcp — Fri, 08 May 2026 10:45:35 +0000

ServiceNow MCP: Automate ITSM workflows without leaving your AI agent

ServiceNow MCP is the official integration that lets Claude, Cursor, and other AI agents interact directly with ServiceNow's ITSM platform. Instead of context-switching between your IDE and the ServiceNow portal, you can now orchestrate incidents, change requests, and approvals entirely through natural language prompts in your agent.

This is particularly useful for ops teams building automation workflows or developers who need to integrate service desk operations into their deployment pipelines. The server speaks both directions—your AI agent can read incident data, search the knowledge base, and write changes back to ServiceNow in real time.

What It Does

ServiceNow MCP unlocks full ITSM lifecycle management through your AI agent. You can create and update incidents, problems, and change requests; search knowledge base articles to surface solutions during troubleshooting; manage approval workflows and SLAs; query asset and configuration item data; and access service desk analytics to identify trends.

The server acts as a direct API bridge, so your agent can handle repetitive ticket routing, auto-assign based on skill tags, fetch related configuration items during incident triage, or even generate reports from trend data. Think of it as giving Claude or Cursor read-write access to your entire service desk—without the browser tab.

How to Install

Run this single command:

npx -y @servicenow/mcp

Then add the server to your Claude Desktop configuration (or Cursor/Windsurf equivalent):

{
  "mcpServers": {
    "servicenow-mcp": {
      "command": "npx -y @servicenow/mcp",
      "env": {
        "SERVICENOW_INSTANCE": "your-instance.service-now.com",
        "SERVICENOW_API_KEY": "your-api-key"
      }
    }
  }
}

Replace SERVICENOW_INSTANCE and SERVICENOW_API_KEY with your credentials. See the official docs for authentication details.

Real-World Use Cases

Incident triage in Cursor: Paste an error log, ask Claude to search the knowledge base for solutions, and auto-create an incident with relevant config items attached—all without leaving your editor.
Change request automation: Describe a deployment, let your AI draft a change request with SLA calculations, and route it to the approval queue based on impact level.
On-call escalation: Build a Slack bot or CLI tool that queries open incidents, fetches analytics, and drafts summaries for your incident commander—all via natural language.

Full install guides for Claude Desktop, Cursor, Windsurf, and more at CuratedMCP.

MCP Ecosystem Week 19: Data Intelligence and Safety Tools Lead New Arrivals

curatedmcp — Mon, 04 May 2026 10:38:54 +0000

MCP Ecosystem Week 19: Data Intelligence and Safety Tools Lead New Arrivals

The MCP marketplace continues to expand with four new servers landing this week, reflecting a clear market demand for data accessibility and content safety. Week 19 brings tools spanning SEO analytics, conversational UI, marketing data aggregation, and threat detection—a diverse mix that signals the ecosystem is moving beyond developer-centric integrations toward enterprise and consumer-facing use cases.

This Week in MCP

Four new verified servers joined the catalog:

DataForSEO MCP Server opens SEO and SERP data to any MCP client. For developers building competitive analysis or content tools, this eliminates the friction of API documentation and token management.

Perspective AI replaces form-based interactions with adaptive AI conversations. The shift from static inputs to natural dialogue is gaining traction—this server makes that pattern pluggable for any agent.

Windsor.ai MCP Server aggregates marketing, sales, and customer data across 325+ sources and exposes it via plain English queries. This is a meaningful consolidation play: instead of agents juggling multiple vendor SDKs, they get a unified interface.

Tuteliq detects grooming, bullying, fraud, and 16+ threat types across text, voice, image, and video. Content moderation and safety are increasingly non-negotiable for production AI systems, and multimodal threat detection filling a real gap.

All four are available free—a continued pattern in the ecosystem.

Trending Servers

The top three trending servers by page views this week reveal what developers are actively integrating:

GitHub Copilot MCP (98K views) leads. The appeal is clear: developers want Copilot's code intelligence—completions, explanations, reviews—available outside the IDE, in their own MCP workflows. It's a portability play.

OpenAI MCP (87K views) maintains second place. Official integrations matter. Having GPT-4o, DALL-E, Whisper, and Embeddings available as MCP resources means developers can route AI workloads through standardized tooling rather than managing OpenAI SDK fragmentation.

Figma MCP (82K views) rounds out the top three. Design-to-code and design system integration are high-friction problems. Direct access to design files, components, and tokens within an AI agent workflow is a legitimate productivity unlock—and the traffic reflects that need is real.

The pattern: developers prioritize integration servers that reduce context switching and vendor lock-in.

Quick Take

The all-free pricing on this week's new arrivals (and 67 of 68 total servers) reflects a strategic choice by vendors: expand adoption fast, own the integration layer, monetize upstream or at the platform level. But there's a secondary story: we're seeing category clustering around data intelligence (SEO, marketing, CRM aggregation) and safety/moderation. These are problems that individual AI developers can't easily solve in-house. The marketplace is becoming a repository of specialized, repeatable capabilities—the opposite of monolithic platforms. That's healthy ecosystem behavior.

Browse all 68 verified servers at CuratedMCP.

Okta MCP: Let Your AI Agent Handle Identity Management

curatedmcp — Sat, 02 May 2026 10:12:28 +0000

Okta MCP: Let Your AI Agent Handle Identity Management

Identity and access management shouldn't require manual ticket shuffling. Okta MCP is the official Okta integration for AI agents—it lets Claude, Cursor, or Windsurf directly manage users, groups, authentication workflows, and security events within your Okta tenant.

If you're managing identity infrastructure, this means your AI can provision users, enforce access policies, investigate authentication failures, and respond to security incidents without you leaving your IDE or chat window.

What It Does

Okta MCP grants AI agents safe, scoped access to your identity workflows. Your agent can:

Manage users at scale: create, update, deactivate accounts, reset passwords, and manage recovery rules
Control access: assign users to groups and applications, manage roles, and inspect sign-on policies
Monitor security: query authentication events, check MFA status, and analyze access patterns
Automate OAuth/OIDC: manage client configurations and token policies

The server is built for security and IT teams who want AI to reduce toil—faster user onboarding, quicker incident response, and easier policy enforcement. Everything is gated by your Okta API token, so permissions stay locked to what you've explicitly authorized.

How to Install

npx -y @okta/mcp

Then add it to your Claude Desktop config:

{
  "mcpServers": {
    "okta-mcp": {
      "command": "npx -y @okta/mcp",
      "env": {
        "OKTA_DOMAIN": "your-org.okta.com",
        "OKTA_API_TOKEN": "your_api_token_here"
      }
    }
  }
}

Grab your API token from Admin → API → Tokens in your Okta console. Full setup guides for Cursor and Windsurf are in the docs.

Real-World Use Cases

Onboarding sprint: Ask Claude to provision 12 new contractors, assign them to the engineering group, activate their Slack app, and send a summary—all in one prompt.
Incident response: When a security alert fires, ask your agent to pull MFA status for a suspicious user, review their recent sign-ons, and deactivate sessions—no context switching.
Access review automation: Have your agent query who has admin roles, who hasn't logged in this quarter, and which legacy OAuth clients are still active—then generate a compliance report.

Full install guides for Claude Desktop, Cursor, Windsurf, and more at CuratedMCP.

IBM Cloud MCP: Control Your IBM Cloud Infrastructure From Claude

curatedmcp — Fri, 01 May 2026 10:06:17 +0000

IBM Cloud MCP: Control Your IBM Cloud Infrastructure From Claude

Managing IBM Cloud resources usually means jumping between dashboards, CLIs, and documentation. IBM Cloud MCP changes that by giving AI agents direct access to your infrastructure, databases, and AI services. Built by IBM for enterprise teams, this official server lets you provision compute, query data, and inspect security policies without leaving your Claude chat or code editor.

What It Does

IBM Cloud MCP is a Model Context Protocol server that bridges Claude, Cursor, and other AI agents to IBM Cloud's full platform. Instead of manual API calls or console clicks, you can ask Claude to spin up Kubernetes clusters, deploy serverless functions, or query Watson AI models directly.

The server handles compute provisioning (Kubernetes, VMs, Cloud Functions), data operations (Cloudant, Db2, Databases for PostgreSQL), and enterprise workflows like identity management and security audits. You get real-time access to logs, monitoring alerts, and metrics across regions. It's designed for teams managing hybrid cloud workloads—IBM Cloud alongside on-premises systems.

For developers, this means faster iteration on cloud-native apps, faster infrastructure troubleshooting, and the ability to automate repetitive cloud tasks through natural conversation.

How to Install

Install via npm:

npx -y @ibm/mcp

Then add it to your Claude Desktop config file (~/Library/Application Support/Claude/claude_desktop_config.json on macOS):

{
  "mcpServers": {
    "ibm-cloud-mcp": {
      "command": "npx -y @ibm/mcp",
      "env": {
        "IBM_CLOUD_API_KEY": "your-api-key-here"
      }
    }
  }
}

Grab your IBM Cloud API key from the IAM dashboard, set it as an environment variable, and restart Claude.

Real-World Use Cases

Deploy and scale an app in seconds: "Deploy my Node.js app to IBM Cloud Kubernetes with 3 replicas across Dallas and London regions, then show me the logs."
Debug database issues with context: "Query my Db2 instance for slow queries in the last hour, then suggest indexes based on the execution plans."
Audit security posture: "List all IAM policies for my service accounts, identify overly permissive roles, and generate a compliance report."

Full install guides for Claude Desktop, Cursor, Windsurf, and more at CuratedMCP.