Manage Cloudflare infrastructure from agents. Deploy Workers, manage DNS records, query analytics, and interact with the full Cloudflare API.

MCP server for managing Docker containers, images, and volumes through natural language commands.

Official AWS Labs MCP server enabling AI assistants to interact with AWS services via CLI commands with IAM-based permissions and CloudTrail logging.

MCP server for interacting with Kubernetes clusters via kubectl, supporting pod management, deployments, and cluster inspection.

Official Vercel MCP server. Manage deployments, projects, domains, environment variables, and logs from Vercel's frontend cloud within agent workflows.

MCP server for Datadog. Query metrics, logs, traces, monitors, and dashboards from Datadog's observability platform within agent workflows.

Official Grafana MCP server. Query dashboards, panels, data sources, and alerts from Grafana for AI-driven observability and incident response workflows.

Official PagerDuty MCP server. Manage incidents, services, on-call schedules, and escalation policies. Acknowledge, resolve, and create incidents directly from AI agents.

Manage Render cloud deployments from AI agents. Deploy services, check build logs, manage environment variables, and monitor service health.

Manage Google Cloud Platform resources: Compute Engine, Cloud Storage, BigQuery, Cloud Run, GKE, and IAM. Deploy, monitor, and scale infrastructure via the GCP API.

Manage Microsoft Azure resources: VMs, App Services, Blob Storage, Azure OpenAI, AKS, and more. Deploy and operate cloud workloads via the Azure Resource Manager API.

Infrastructure as Code using general-purpose programming languages. Deploy and manage cloud resources across AWS, Azure, GCP, and Kubernetes via Pulumi's API and state management.

MCP server for Prometheus metrics. Run PromQL queries, inspect targets, and retrieve time-series metric data for AI-driven infrastructure monitoring.

MCP server for OpenTelemetry. Query traces, metrics, and logs from OTEL-compatible backends (Jaeger, Zipkin, OTLP) for AI-driven observability workflows.

Fly.io infrastructure management via MCP. Deploy apps, manage machines, inspect logs, and scale services globally from agents.

Netlify deployment and site management via MCP. Trigger deploys, manage environment variables, inspect build logs, and control DNS from agents.

Railway infrastructure management via MCP. Deploy services, manage environment variables, inspect deployments, and scale projects from agents.

Official Grafana MCP server for Loki log querying. Execute LogQL queries against Loki instances with time range and multi-tenant support.

Official Axiom MCP server. Query logs, traces, and event data using APL. List, create, and inspect datasets with full schema visibility.

Official Portainer MCP server. Manage Docker and Kubernetes environments, stacks, users, and teams through AI assistants with 40+ tools.

MCP server for Argo CD. List, create, sync, and manage GitOps applications and resources. View logs and run resource actions.

MCP server wrapping the Backstage Catalog API. Browse, filter, and create entities in your Backstage developer portal from AI assistants.

Talk to any Cloudflare Worker from MCP clients. Bridge between MCP protocol and Worker endpoints for serverless tool hosting.

Unified MCP server for querying OpenTelemetry traces across backends (Jaeger, Tempo, Traceloop). Enables AI agents to analyze distributed traces for debugging.

Security scanning for AI infrastructure and MCP servers. Detect shadow IT, credential exposure, tool poisoning, and supply chain risks in agent deployments.

AI-powered security operations for Wazuh SIEM. Enables conversational SOC workflows for threat detection, incident triage, compliance checks, and real-time monitoring.

MCP server for the VirusTotal API providing security analysis of URLs, files, IPs, and domains with relationship analysis and pagination support.

Unified threat intelligence MCP server aggregating data from AlienVault OTX, AbuseIPDB, GreyNoise, and abuse.ch feeds for correlated IOC lookups.

Official Google MCP server for threat intelligence (VirusTotal/GTI), Security Operations (Chronicle), SOAR, and Security Command Center integration.

Security scanner for MCP servers using YARA, LLM-as-judge, and Cisco AI Defense engines to detect prompt injection, tool poisoning, and supply chain threats.

MCP server integrating with MISP (Malware Information Sharing Platform) for threat intelligence including IOC search, threat actor lookup, and malware detection.

MCP server for ProjectDiscovery Nuclei vulnerability scanner enabling context-aware scanning with intelligent template selection for CVEs, exposures, and misconfigurations.

Interact with Jenkins CI/CD pipelines via natural language. Trigger builds, check job status, retrieve logs, manage parameters, and monitor pipeline health.

Manage RabbitMQ message queues via natural language. Publish and consume messages, inspect queue depths, manage exchanges and bindings, and monitor broker health.

Interact with HashiCorp Consul for service discovery and configuration. Query service catalog, manage KV store, inspect health checks, and control access policies.

Trigger, monitor, and manage Apache Airflow DAGs via MCP. Agents can start pipeline runs, check task status, read logs, and respond to failures without human intervention.

Orchestrate and observe Prefect workflows via MCP. Trigger flow runs, monitor states, manage deployments, and retrieve logs — letting agents drive data pipeline operations end to end.

Run serverless GPU and CPU workloads on Modal via MCP. Deploy functions, trigger jobs, and retrieve results — giving agents access to scalable compute without managing infrastructure.

Access and manage secrets from Doppler via MCP. Agents can fetch environment-specific configs, rotate secrets, and audit access — keeping credentials out of code and out of prompts.

Interact with MinIO S3-compatible object storage via MCP. Upload, download, list, and delete objects across buckets — giving agents a self-hosted storage layer with full S3 API compatibility.

Google-managed MCP server for Cloud Run serverless container platform. Agents can deploy services, manage revisions, configure traffic splitting, and monitor running containers. IAM-authenticated with full audit logging.

Official Dynatrace MCP server for observability. Query metrics with DQL, investigate problems and vulnerabilities, analyze Kubernetes events, forecast timeseries data, and access production insights from AI agents.

MCP server for Temporal workflow orchestration. Lets AI agents start, monitor, signal, cancel, and inspect workflows entirely through natural language. Exposes 19 tools covering the full Temporal surface area including batch operations and schedule management via cron expressions.

MCP server for Microsoft Purview compliance and governance. Connects AI agents to Purview's compliance engine for audit trails, retention policies, data governance workflows, and eDiscovery. Transforms complex compliance operations into conversational commands.

Official Home Assistant MCP server integration for controlling smart home devices, automations, and scenes via AI assistants.

Official IBM QRadar SIEM MCP server. Search offenses, run AQL queries, manage reference sets, and investigate security incidents through AI agents.

MCP server for IBM i systems. Enables AI agents to interact with IBM i through SQL-based tools for performance monitoring, security auditing, and database administration.

MCP server for Trigger.dev — the open-source background jobs and workflow platform. Trigger jobs, monitor runs, inspect logs, and manage scheduled tasks from AI agents without polling or webhooks.

MCP server for LaunchDarkly — the feature management platform. Query and update feature flags, manage targeting rules, roll out features progressively, and automate flag lifecycle from AI agents.

MCP server for Statsig — the feature flagging and experimentation platform built for AI-native companies. Run A/B tests, manage feature gates, query experiment results, and automate rollout decisions from agents.

MCP server for Logfire — Pydantic's OpenTelemetry-based observability platform. Query traces, logs, and spans from Python and AI applications, inspect errors, and monitor agent runtime behavior in real time.

MCP server for Drata — the compliance automation platform for SOC 2, ISO 27001, HIPAA, and GDPR. Query controls, monitor evidence collection, track audit readiness, and automate compliance workflows from AI agents.

Official unified AWS Agent Toolkit MCP server (GA May 2026). A single tool to execute any of 15,000+ AWS API operations with IAM-based guardrails, CloudWatch metrics, CloudTrail logging, and sandboxed Python execution for multi-step workflows.

Managed MCP server from AWS that gives AI agents secure, auditable access to 15,000+ AWS API operations with IAM guardrails, CloudWatch metrics, and CloudTrail logging.

Official AWS managed MCP server giving AI coding agents secure, auditable access to AWS services. Supports infrastructure provisioning, CloudWatch queries, S3 operations, Lambda management, and over 40 AWS service integrations with IAM-scoped permissions.

Official Oracle Cloud Infrastructure MCP server for AI-driven data protection and disaster recovery. Gives AI agents access to backup policies, recovery points, and resilience status across OCI workloads with full audit logging.

MCP server from Digi International giving AI agents access to IoT device fleet management through Digi Remote Manager and Genesis platforms. Enables natural language queries across device fleets, automated workflows, and configuration insights.

Open-source computer-using agent framework that turns local desktop automation into a standard MCP service. Supports browser automation, terminal access, desktop control via computer vision, and multi-agent orchestration across macOS, Linux, and Windows. Achieves 82% on the OSWorld benchmark.

Open-source MCP registry and proxy from IBM that federates tools, agents, and APIs into one clean endpoint for AI clients. Provides a secure, intelligent gateway between AI agents and the tools they use, simplifying MCP server management with built-in discovery, routing, and access control.

Enterprise agent control plane from Trust3 AI that governs and secures MCP connections and agent-to-agent communications. Provides identity access management for AI agents, content firewalling, single-purpose credential isolation, and immutable audit trails for litigation-grade compliance across agentic AI workloads.

Open-source MCP governance extensions from Microsoft for .NET that add rate limiting, content filtering, guardrails, and audit logging to MCP server deployments. Acts as a governance proxy layer between MCP clients and downstream servers, enforcing enterprise policies on tool usage, input validation, and output filtering for agentic AI workloads.

MCP server from Boomi for governing and managing AI agent tool access across the enterprise. Provides a vendor-agnostic catalog of MCP servers instantly available to agents, with centralized governance, access controls, and integration with the Boomi platform's automation and integration capabilities.

Open-source MCP server monitoring toolkit from BlueRock for Python applications. Provides hooks that observe tool calls, session events, module imports, and subprocess activity in Pythonic MCP server deployments. Designed for security teams and platform engineers to gain visibility into MCP server behavior in production.

Official MCP server from Splunk for integrating AI agents with the Splunk observability and security platform. Version 1.1 enables AI assistants to search logs, create alerts, query dashboards, and perform security investigations through natural language. Supports SPL query generation and real-time data access for incident response workflows.

MCP server from ArmorCode providing contextual application security posture management (ASPM) data for AI-powered automation. Part of the Anya Agents framework, it enables AI agents to access vulnerability findings, risk prioritization, and remediation guidance across the software development lifecycle.

Zero trust MCP architecture from Versa Networks that acts as a secure broker between AI agents and enterprise infrastructure. Validates every AI agent action against security policies before execution, applying identity verification, least-privilege access controls, and real-time threat inspection to MCP workflows. Extends zero trust networking principles to agentic AI communications.

Open-source MCP server from AWS Labs for comprehensive threat modeling with automatic code validation. Generates threat models from architecture descriptions, validates threats against actual source code, and produces security recommendations. Supports STRIDE methodology, data flow analysis, and automated mitigation suggestions for cloud-native applications.

Built-in MCP server for the Dagu workflow engine, a lightweight single-binary orchestrator with a Web UI. Enables AI agents to create, update, and run DAG-based workflows with full traceability. Supports declarative YAML definitions, parallel execution, conditional branching, and retries across Linux, macOS, and Windows.

Comprehensive MCP server exposing the complete Zabbix monitoring API with 237 tools. Supports multi-server connections, OAuth 2.1 and bearer token authentication, PDF report generation, and systemd deployment. Compatible with ChatGPT, Claude, VS Code, and other MCP hosts for full infrastructure monitoring through natural language.

Secure credential management MCP server for terminal users and AI agents. Stores and retrieves sensitive secrets with modern encryption and built-in MCP server support. Provides agents with safe access to API keys, tokens, and passwords without exposing raw credentials in prompts or logs.

Infrastructure management MCP server embedded in SUSE Rancher Prime and SUSE Multi-Linux Manager. Enables AI agents to monitor, troubleshoot, and optimize Linux servers and Kubernetes clusters. Includes AI engineering skills for identifying system faults, correlating logs, and submitting patches or service restarts through autonomous workflows.

Security testing MCP server from Detectify that brings vulnerability scanning and attack surface data into AI-driven development workflows. AI agents can receive remediation tasks, generate patches, trigger validation scans, and present verified fixes for human review. Supports conversational querying of scan results, asset monitoring, and severity-based finding prioritization.

Security operations MCP server from Vectra AI that gives AI assistants standardized access to threat detection and investigation capabilities. Enables security analysts to investigate incidents, reconstruct attack timelines, and generate context-rich reports through natural language prompts in Claude Desktop, Cursor, and other MCP clients. Requires no changes to existing Vectra deployments.

First production-ready MCP server for certificate lifecycle management from Sectigo. Enables AI agents and assistants to discover, monitor, renew, and manage SSL/TLS certificates across enterprise environments. Designed for security and IT operations teams adopting agentic AI, providing programmatic access to Sectigo Certificate Manager capabilities through standardized MCP interfaces.

Security governance platform from Noma that discovers, governs, and enforces access policies for AI agents and MCP servers across the enterprise. Assigns each autonomous agent a distinct, attributable identity when connecting to MCP servers and tools. Provides visibility into which agents access which tools, enforces least-privilege policies, and enables security teams to monitor agentic AI behavior at scale.

Content provenance and verification MCP server from Digimarc that enables AI agents and orchestration frameworks to stamp, verify, log, audit, and retrieve provenance information through MCP-compatible interfaces. Grounded in the C2PA standard adopted by Adobe, Google, Microsoft, and OpenAI. Every provenance seal is policy-gated, issued only when agent identity, artifact integrity, and request timing satisfy defined trust criteria. Backed by Digimarc's Illuminate platform with optional watermarking for durability when metadata is stripped during processing.

Open-source compliance automation MCP server from Comp AI that connects AI agents to compliance frameworks, evidence collection, and audit workflows. Enables AI assistants to query compliance status, manage evidence artifacts, and automate control assessments across SOC 2, ISO 27001, and other frameworks. Part of the Comp open-source compliance platform designed for startups and growing companies.

MCP server from Infoblox that makes enterprise network, security, and asset intelligence available to third-party AI assistants and agents through the Model Context Protocol. Part of the Infoblox IQ agentic operations platform announced June 2026. Provides access to DNS queries, DHCP leases, IP address assignments, device activity, and security events flowing through the Infoblox Platform. Built on 25 years of DDI operational data from thousands of enterprise deployments including the majority of Fortune 100 companies.

MCP server from N-able that connects AI assistants including ChatGPT, Claude, and Microsoft Copilot to live Unified Endpoint Management data. Enables IT teams to query systems, check device health, and take controlled management actions in real time through natural language. Part of N-able's AI-native UEM platform targeting up to 70% faster IT operations for managed service providers and enterprise IT departments.

Enterprise zero-trust MCP gateway from NetFoundry that provides secure access to MCP servers without exposing them to the network. Built on Identity-First Reachability with no open inbound ports, no VPNs, and no firewall changes. Supports multi-backend aggregation, tool namespacing, structural permission filtering, per-client session isolation, centralized multi-user management, and role-based access control. Denied tools are removed from the registry entirely rather than checked at runtime. Paired with an LLM Gateway supporting semantic routing across OpenAI, Anthropic, Azure, Bedrock, Vertex AI, and Ollama with built-in PII detection and prompt injection filtering.

MCP server from New Relic that provides standardized access to observability data for AI agents and assistants. Integrates New Relic telemetry including application performance monitoring, infrastructure metrics, logs, distributed traces, error tracking, and alerting data into MCP-compatible AI tools. Enables AI agents to query production health, investigate incidents, analyze performance trends, and correlate signals across the full stack without manual dashboard navigation. Available in public preview for Claude Desktop, Cursor, and other MCP clients.

MCP server from Kajeet that connects AI agents to the SentinelOS IoT device management platform. Launched June 2026 as part of the SentinelOS AI agent-ready operating layer. Enables approved AI agents and assistants to query device status, connectivity health, policy data, and network diagnostics under governed access with authentication and role-based permissions. Part of a broader platform including the Sentinel CLI, REST APIs, and webhooks for managing connected device fleets across enterprise and education deployments.

MCP server from Domotz that connects AI assistants to the Domotz IT network management and monitoring platform. Launched June 2026. Enables AI agents to check device quotas and allowances, query historical metrics with trend analysis, and monitor network device inventory directly through MCP-compatible tools. Designed for MSPs and IT teams managing distributed network infrastructure who want to query device and network health through natural language.

Platform from Nexla that enables enterprise teams to build governed, task-specific MCP servers through conversation without writing integration code. Connects to 600+ enterprise systems and 10,000+ available tools and actions spanning Salesforce, Snowflake, SAP, NetSuite, Workday, Databricks, ServiceNow, and more. Users describe a business outcome, provide system access, and Nexla autonomously discovers data and permissions, selects minimum required tools, assembles cross-system context, and generates a production-ready MCP server. Servers inherit access controls, credential management, audit logging, and data lineage. Currently in Early Access. Announced June 9, 2026.

Open-source MCP gateway from Lunar.dev that provides a single governed entry point for all agent-to-tool interactions across an organization. Enforces access control, auditability, and policy across every MCP server with approximately 4ms p99 latency. Features include identity-based governance, OAuth passthrough so each end user authenticates with upstream services under their own credentials, and integration with Cursor, Claude Desktop, Claude Code, VS Code, and any MCP-compatible client. Licensed under MIT for non-production use with an Enterprise tier for hosted deployment, automated risk scoring, and additional governance. Deployed by Fortune 200 enterprises to govern AI adoption across engineering teams.

Platform from Inconvo (YC S23) that enables SaaS companies to launch remote authenticated MCP servers for their customers in minutes without building custom infrastructure. Connect an existing database and authentication provider to deploy an AI reporting agent attached to customer data that handles conversation state, schema introspection, agent observability, multi-tenancy, and query safety as managed services. Customers can link the MCP server to ChatGPT and other AI tools to query their SaaS data through natural language. Includes agent traces and logs for out-of-the-box observability and a semantic layer for improving agent performance through context engineering.

Managed MCP server platform from Truto that auto-generates MCP servers for over 250 SaaS integrations. Every connected account automatically becomes a set of tools that any MCP-compatible client can invoke. Uses real-time pass-through architecture with no data caching for always-fresh results. Each MCP connection is scoped to a specific integrated account ensuring agent actions are bounded to the correct data. Compatible with Claude, ChatGPT, Cursor, and any MCP client. Supports CRM, HRIS, ATS, accounting, ticketing, and calendar integrations.

MCP server from AppOmni that enables AI agents and security platforms to access deep SaaS security intelligence. AskOmni operates as an MCP server providing insights into SaaS identities, security posture, data exposures, and user behaviors for threat investigations across the SaaS domain. Integrates with SIEM, NDR, XDR, and IAM solutions to enrich SaaS alert context, detect real threats, and filter false positives. Can identify and block unauthorized prompt activity in ServiceNow agentic AI. Announced at RSA Conference. Works with Microsoft Copilot, Claude, and other MCP-compatible security tools.

Enterprise MCP gateway from Cequence that transforms any API into MCP-compatible tools with automated security, governance, and DLP scanning. Supports over 140 verified enterprise application integrations across collaboration, DevOps, marketing, and security platforms. Agent Personas feature provides infrastructure-level privilege scoping for autonomous AI agents with least-privilege access controls. Applies DLP scanning with 100+ detection types covering PII, credentials, financial data, and health records to both agent requests and MCP server responses. Deploys as SaaS or in customer private cloud. Works with Claude, Cursor, and Windsurf IDE.

Official unified MCP server from Microsoft that enables AI agents to securely access and manage Azure resources through natural language. Provides tools for Azure Resource Graph queries, ARM template deployments, Azure Kubernetes Service cluster management, Azure Container Apps, Cosmos DB, and AI Foundry. Built into Visual Studio 2026 and available as a standalone server. Inherits the local machine active az CLI session credentials for enterprise-grade security with no static connection strings. Open source on GitHub under the microsoft/mcp catalog.

Hosted MCP server from Honeycomb that lets AI agents directly query, analyze, and visualize observability data including traces, triggers, and SLOs using natural language. First observability vendor to treat MCP as a fully managed service rather than a self-hosted tool. Available in the AWS Marketplace AI Agents and Tools category. Expanded in March 2026 with MCP capabilities for Claude Code, Cursor, AWS DevOps Agent, and additional platforms. Includes Agent Skills for guided observability workflows. May 2026 introduced Agent Timeline, Canvas Agent, and Canvas Skills for AI agents in production.

Cloud-native MCP server from Fastio providing AI agents with persistent file storage, memory, and semantic search capabilities. Exposes 19 tools via Streamable HTTP and SSE transports covering file CRUD operations, folder management, and RAG-powered semantic search. Intelligence Mode auto-indexes files for retrieval-augmented generation. File locks prevent conflicts when multiple agents access the same file. Agents get dedicated accounts with 50GB free storage and 5000 monthly credits. Compatible with Claude Desktop, Cursor, Cline, and any MCP client that speaks HTTP.

MCP traffic monitoring module from NGINX that inspects Model Context Protocol traffic in real time and reports on throughput, latencies, errors, and traces. Introduces agentic observability at the infrastructure layer by providing visibility into MCP server interactions as they happen. Enables operators to monitor AI agent to MCP server communication patterns, identify bottlenecks, and troubleshoot latency issues. Integrates with existing NGINX deployments as a module. Announced on the NGINX Community Blog as part of the agentic AI infrastructure stack.

MCP server built into SafeDep vet that protects AI-generated code from slopsquatting attacks, malicious packages, known vulnerabilities, and other software supply chain risks. When an AI coding assistant generates dependency recommendations, vet's MCP tools validate packages against vulnerability databases, malware signatures, and package registry metadata in real time. Supports npm, PyPI, Go, Maven, and other major ecosystems. Open source on GitHub. Designed for developers using AI coding assistants who need automated dependency safety checks.

Official MCP server from CrowdStrike that connects AI agents to the Falcon cybersecurity platform for intelligent security analysis. Provides tools for querying detections, investigating hosts, searching indicators of compromise, and analyzing threat intelligence through the CrowdStrike Falcon API. Supports host containment actions, RTR session management, and vulnerability assessment. Open source on GitHub under the CrowdStrike organization. Currently at v0.12.0 with active development. Works with Claude, Cursor, and any MCP-compatible client.

Community MCP server for Automox, the cloud-native IT endpoint management platform. Acts as a stateless proxy between AI assistants and the Automox API, enabling device management, patch deployment, software installation, and compliance reporting through natural language. Supports querying device inventory, triggering patch scans, managing server groups, and executing Worklets. Open source on GitHub under AutomoxCommunity. Works with Claude Desktop, Cursor, and any MCP-compatible client. Designed for IT operations teams managing endpoint fleets.

MCP server that provides 37 tools for managing multiple SSH connections through AI agents. Enables Claude Code, OpenAI Codex, and other MCP-compatible clients to execute remote commands, transfer files, manage tunnels, and handle multi-server orchestration through natural language. Supports concurrent connections to multiple hosts with session persistence. Includes tools for command execution, file upload and download, port forwarding, and server inventory management. Open source on GitHub under bvisible. Designed for DevOps teams and system administrators managing remote server fleets.

Open-source MCP gateway that routes requests across multiple MCP servers with 0.02ms overhead and 46K+ requests-per-second cache hits. Provides authentication, guardrails, plugin support, and multi-server routing in a single gateway layer. Validated with 1,341 tests. Enables organizations to consolidate access to multiple MCP servers behind a single entry point with policy enforcement, caching, and observability. Designed for production deployments where multiple MCP servers need unified access control and traffic management. Created by lacausecrypto.

Open-source security governance framework from Cisco AI Defense for agentic AI systems. Scans AI agent skills and MCP servers for vulnerabilities using a combination of the Cisco AI Defense inspect API, YARA rules, and LLM-as-a-judge to detect malicious MCP tools, prompt injections, tool poisoning, and toxic flows. Includes Skills Scanner, MCP Scanner, AI Bill of Materials generator, and CodeGuard for static analysis. Ensures every agent skill is scanned, every MCP server verified, and every AI asset cataloged automatically. Released at RSA 2026. Apache 2.0 licensed. Python-based.

Security scanner from Snyk for AI agents, MCP servers, and agent skills. Part of Snyk's Agent Security suite launched at RSAC 2026. Scans for common security vulnerabilities including prompt injections, tool poisoning, and toxic flows in agent skills and MCP servers. Provides MCP server governance through automated scanning and compliance checks. Available as open preview. Integrates with Snyk Studio for AI-assisted vulnerability detection and remediation in coding workflows. Works alongside Agent Guard for real-time enforcement and agent red-teaming via CLI.

Self-hosted MCP gateway and registry for managing multiple MCP servers from a single endpoint. Built in Go with centralized tool discovery, access control, and multi-transport support for both Streamable HTTP and stdio. Agents connect to one gateway instead of multiple endpoints. Supports enterprise workflows with authentication, observability, and registry metadata. Deploys via Docker with a Postgres backend. Works with Claude Desktop, Cursor, and any MCP-compatible client. 1.1K GitHub stars. Open source under MIT license.

Unified middleware MCP server with a web GUI for managing multiple MCP server connections. Acts as a proxy that dynamically aggregates MCP servers into a single unified endpoint. Supports multi-workspace switching for isolating contexts between different databases or projects. Self-hostable via Docker Compose. Works with any MCP client including Claude Desktop, Cursor, and VS Code. 2.4K GitHub stars. Free and open source. Created by metatool-ai.

Unified MCP runtime that aggregates multiple MCP servers behind a single serve process. Provides CLI mode with progressive discovery commands for inspect, instructions, and run workflows. Supports template servers for per-client or per-session resolution, async and lazy loading for faster startup, and namespace isolation. Built in TypeScript. Works with Claude Code, Cursor, Codex, and other MCP-compatible agents. Solves configuration sprawl by centralizing MCP server management. Created by 1mcp-app.

Reverse proxy and management layer for MCP servers from Microsoft enabling scalable, session-aware routing and lifecycle management in Kubernetes environments. Provides a data gateway with session affinity, control plane for deploying and managing MCP servers, and a management portal with in-browser JSON-RPC test console. Includes enterprise-ready telemetry, access control, and observability. Open source on GitHub. Built in C#.

Google Cloud API governance layer that exposes existing APIs cataloged in Apigee API hub as MCP tools for AI agents. Provides 30+ built-in policies for authorization, authentication, and governance without requiring code changes or new MCP server deployments. Supports OAuth 2.1 and OIDC authentication, tool-level analytics, and MCP-specific traffic segmentation. Transcodes existing API specifications automatically. Commercial via Google Cloud. Created by Google.

Open source observability platform MCP server that connects AI agents to logs, metrics, traces, and real user monitoring data. Enables natural language queries instead of SQL or PromQL, automated alert creation and stream management, and AI-powered analysis for error pattern detection and latency spike diagnosis. Unifies all observability data into a single conversational interface. 140x lower storage costs than alternatives. Open source. Created by OpenObserve.

IoT device management MCP server from Digi International that connects AI assistants to Digi Remote Manager and Genesis platforms. Enables natural language interfaces to query device fleets, automate workflows, generate configuration insights, and monitor connected devices across enterprise IoT deployments. Supports device health monitoring, fleet analytics, and automated troubleshooting. Commercial with managed hosting. Created by Digi International.

37-tool MCP server for remote SSH management. Execute commands, transfer files, manage databases, create backups, monitor server health, and automate DevOps tasks across multiple servers from Claude Code or Codex.

Unofficial MCP server for Home Assistant smart home control. Lets AI assistants interact with Home Assistant to manage devices, automations, scenes, and monitor sensor data through natural language.

Official ioBroker MCP server adapter for smart home automation. Exposes ioBroker installations to MCP-capable AI clients via Streamable HTTP transport, enabling reading and controlling smart home devices, running diagnostics, and searching adapter repositories.

Enterprise integration platform MCP server from SnapLogic enabling teams to expose existing integrations, data flows, and workflows as MCP servers. Includes a MCP Server Pipeline Builder that walks through creating a functional MCP server pipeline in steps, allowing any SnapLogic integration to become AI-agent accessible.

Managed cloud MCP server from ClickHouse bringing ClickStack observability investigation capabilities to external AI agents. Enables Claude, Cursor, and Codex to query observability data, run diagnostic investigations, and analyze performance traces directly through the ClickHouse Cloud platform without local server setup.

MCP server from LucidLink exposing its distributed streaming file system to AI agents. Gives multi-agent systems a persistent, writable file layer with shared state across clouds, on-premises, and edge environments. Supports block-level streaming, global file locking, and zero-knowledge AES-256 encryption so agents, applications, and humans work from the same files without copying or moving data.

MCP server for ESPHome providing AI assistants with direct access to ESPHome device management and the new Device Builder functionality introduced in ESPHome 2026.6. Enables creating, editing, and managing ESPHome device configurations through natural language. Built for the Home Assistant ecosystem, supports the latest ESPHome API methods.

Real-time enforcement gateway from Vorlon for securing AI agent interactions with MCP servers, SaaS apps, and data stores. Announced June 30, 2026. Sits at the protocol layer between AI agents and connected systems to block policy violations, mask sensitive data in transit, and enforce read-only access before transactions complete. Covers agent-specific threats including indirect prompt injection, credential abuse, OAuth token abuse, and agent-to-agent manipulation. Works with any app or data store that has an API or MCP server.

MCP server from Pentera, the exposure validation and offensive security platform, announced June 16, 2026. Enables AI assistants and agents to trigger Pentera security tests, access validated attack findings, and pull remediation guidance directly within AI-powered SecOps workflows. Security teams can validate whether specific vulnerabilities represent real exploitable exposure and prioritize fixes based on proven cyber risk. Integrates with any MCP-compatible environment.

Agentic Development Security platform from Snyk, announced June 24, 2026. Goes beyond the existing Snyk MCP server by governing AI agents in real time, securing MCP server connections, monitoring agent behavior, and scanning AI-generated code across the full development lifecycle. Provides runtime governance for agentic coding workflows in tools like Claude Code, Cursor, and GitHub Copilot. Enforces security policies on which MCP servers agents can connect to and what actions they can take.

Open-source MCP server from FINOS (Fintech Open Source Foundation) for operationalizing the AI Governance Framework inside AI agents, announced June 22, 2026. Supplies structured governance context to AI agents operating in financial services, enabling them to meet industry-specific risk assessment, compliance, and audit requirements. Designed for banks, asset managers, and financial institutions deploying agentic AI systems.

MCP server from Facilio, the connected CMMS platform, announced July 1, 2026 as part of the AI-native evolution of their facility management platform. Makes CMMS data and actions reachable from Claude, ChatGPT, or Gemini, enabling facilities teams to manage operations through AI interfaces instead of traditional dashboards. Supports service request handling, vendor coordination, invoice validation, compliance tracking, and report generation through AI agents.

MCP server from Pressable for managed WordPress hosting operations, with a major expansion shipped June 30, 2026. Enables AI assistants to manage WordPress sites through natural language including cache management, backups, WP-CLI commands, SSL certificates, environment transfers, and password resets. Supports running commands across individual sites or entire portfolios simultaneously. Includes team permission controls for governing who can trigger operational commands through AI agents.

Enterprise MCP server creation tool from SnapLogic, made generally available July 1, 2026. Automatically generates ready-to-run MCP servers from existing integration pipelines, OpenAPI specifications, and API Management services through templates. Organizations can publish MCP tools without rebuilding workflows, writing code, or manually constructing MCP implementations. Designed for enterprise AI adoption at scale with governance controls built in.

MCP server from Gravitee for API management and AI agent governance. Converts existing REST APIs into MCP-compatible tools automatically from OpenAPI specifications without writing code. Includes three capabilities: MCP Tool Server for turning APIs into agent tools, MCP Proxy for governing upstream MCP servers with OAuth 2.1, and APIM MCP Server for managing the Gravitee platform via natural language. Part of Gravitee's AI Agent Management platform with centralized control of agent traffic across MCP and A2A protocols.

AI control plane from Speakeasy for generating, deploying, and governing MCP servers. Generates production-ready MCP servers from OpenAPI specifications through the Speakeasy CLI. Provides a centralized gateway for organizations rolling out MCP across their org with enterprise-grade OAuth 2.1 authentication, API key and bearer token support. Servers can be deployed to Cloudflare Workers, Docker, or any custom infrastructure. Includes full MCP release notes tracking and spec compliance tooling.

MCP server from DataDome for securing MCP infrastructure against malicious AI agents. Provides real-time protection against bot attacks, prompt injection, data exfiltration, agentic denial of service, session hijacking, and tool poisoning targeting MCP servers. Integrates via standard server-side module starting with Node.js and AWS CloudFront Lambda@Edge. Includes native FastMCP integration for Python-based MCP servers. Establishes trust scoring for agentic traffic with autopilot protection at the edge.

Enterprise MCP security and governance platform from MCPTotal with a hub-and-gateway architecture. The Secure Hub provides centralized hosting, credential vaulting, and a vetted catalog of hundreds of curated MCP servers that employees can connect to AI models like Claude and Cursor. The Gateway acts as an AI-native firewall that monitors MCP traffic in real time and enforces access policies. Supports instant zero-conf server deployment, authentication management, and compliance controls for organizations adopting MCP at scale. Founded by security experts Gil Dabah and Dr. Ariel Shiftan.

Official collection of MCP servers from IBM for IBM Cloud services, announced at Think 2026. Built directly into the IBM Cloud CLI native binaries with experimental MCP features. Includes servers for IBM Cloud Logs, Cloud Object Storage, watsonx.ai, and core IBM Cloud platform operations. Enables AI assistants to interact with IBM Cloud infrastructure, query services, manage resources, and access platform capabilities through standardized MCP tools. Part of IBM's broader agentic AI strategy with watsonx.

Community-driven MCP server for managing Akamai's edge platform through natural language. Provides tools for Property Manager configuration, Edge DNS zone and record management, certificate provisioning via CPS, WAF policy management, cache purging by URL or CP code, and reporting. Supports multi-customer Akamai account management via EdgeGrid authentication. Built with TypeScript with modular architecture and comprehensive testing. Compatible with Claude Desktop, Cursor, and VS Code.

Open-source transparent proxy and terminal UI for debugging MCP traffic between AI clients and MCP servers. Described as Wireshark for MCP, it inserts into the data path and shows every JSON-RPC call, response, notification, and server stderr live in a TUI. Features include latency tracking, error highlighting, request replay against fresh server copies, stream filtering by tool, status, direction, and kind, search within frames, and a live timer on pending calls. Single-binary Go tool requiring zero configuration changes to servers. MIT licensed.

Open-source MCP server from Veeam that brings backup, recovery, malware detection, and compliance information into AI-driven operational workflows. Deployed locally as a Docker container with full customer governance. Integrates with Veeam Backup and Replication, Veeam ONE, and Veeam Service Provider Console. Provides a conversational interface for day-to-day operations, planned infrastructure changes, and incident response. No raw operational data is sent to external services unless explicitly chosen. Available on GitHub under open-source license.

Fully managed remote MCP server from Google Cloud that securely connects external AI agents to resources inside a Google Cloud environment. Acts as a bridge between external development tools like Claude Code and Antigravity CLI and Google Cloud architecture. Agents can call models from Model Garden, pull shared prompt templates, and manage Notebooks directly within a project. Protected by Cloud IAM Deny policies ensuring external frameworks only interact with authorized resources. Automatically enabled when the Gemini Enterprise Agent Platform API is enabled. Launched June 30, 2026.

Hosted platform that turns any REST, GraphQL, SaaS, or internal API into a production remote MCP server in minutes. Build visually from a dashboard or let an AI agent create, test, and deploy tools from API docs. Supports OAuth2 Authorization Code, Client Credentials, Bearer token, API key, and Basic Auth for upstream APIs. Includes workflow tools that compose multiple API calls into a single MCP tool, response shaping for cleaner agent output, encrypted credential storage, forkable snapshots, and usage tracking. End users connect live MCP servers to Claude, ChatGPT, Codex, Cursor, VS Code, or custom agents.

Enterprise MCP gateway from Zuplo that sits in front of multiple MCP servers, adding OAuth authorization, upstream credential brokering, tool curation, governance, and observability through a single governed endpoint. Ships a complete OAuth 2.0 authorization server written to the MCP spec with PKCE S256 enforcement. Supports per-user credential brokering where end-users complete upstream OAuth flows while the gateway refreshes tokens per session with per-user revocation and audit logging. Public beta since June 2, 2026.

Enterprise-grade MCP gateway and governance platform that provides centralized security, observability, authentication, and compliance controls for AI agents connecting to internal systems. Features Virtual MCP Bundles for team-specific endpoints exposing minimum required tools, Agent Bundles with M2M auth and scoped tools, one-click deployment transforming local STDIO servers into production services with automatic OAuth wrapping, native Snowflake and Elasticsearch integrations, and SCIM-driven membership with rule-based policy. SOC 2 Type II audited.

Enterprise MCP gateway from Kong Inc. that standardizes, secures, and observes AI agent workflows at scale. Part of the Kong AI Gateway 3.14+ release which supports LLM traffic management, MCP server connectivity, API management, and agent-to-agent (A2A) communication in a single gateway. Includes Kong MCP Registry for discovering and governing MCP servers within Kong Konnect, federated SSO, OAuth 2.0, multi-region failover, and comprehensive audit logging. Announced February 2026 with Agent Gateway capabilities added April 2026.

Open-source SDK and cloud platform for building and deploying MCP servers, clients, and agents. The mcp-use SDK provides a CLI, dev server, Inspector, and managed deploys for spinning up and aggregating MCP servers through a single endpoint. Supports TypeScript and Python with integrations for LangChain, OpenAI, and Anthropic. The cloud platform adds hosted MCP server deployment, authentication, and scaling. YC S25 backed with over 6,000 GitHub stars and 120,000 downloads. Used by teams at NASA, NVIDIA, and SAP.

Agent-integration platform providing 1,000+ managed MCP integrations across 500+ apps with built-in OAuth handling, automatic token refresh, and RBAC controls. Features a Tool Router for dynamic tool discovery through a single endpoint, sandboxed execution, and a centralized MCP Gateway for auth, observability, rate limiting, and governance. Integrates natively with OpenAI, Anthropic, LangChain, AutoGen, and Claude Agent SDK. SOC2/ISO certified. Free tier includes 20,000 tool calls per month.

Enterprise AI gateway and MCP control plane that manages access, discovery, and orchestration of MCP servers across organizations. Features Virtual MCP Server abstraction for aggregating tools into logical managed endpoints, OAuth 2.0 Identity Injection for on-behalf-of user actions, federated SSO, dynamic discovery, multi-region failover, and end-to-end tracing with metadata tagging for compliance. Supports the 2026-07-28 spec including MCP Apps and Tasks extensions for interactive UIs and long-running workflows.

Managed MCP server from Rafay Systems that gives platform, DevOps, and SRE teams a governed interface for querying operational infrastructure context through AI assistants. Connects MCP-compatible clients to the Rafay Platform using existing RBAC and project boundaries. Initial workflows cover fleet intelligence and cost attribution for Kubernetes environments, plus incident diagnosis with structured root-cause analysis for common issues like image-pull errors, crash loops, and failed deployments. Read-only access enforced through Rafay API key authentication. Available via feature flag. Announced July 15, 2026.

MCP server from Primo that connects AI assistants to IT operations for fleet-wide device querying and management. Enables IT teams to ask natural language questions about their entire device fleet, including device status, assignment, last activity, open tickets, and installed software. Provides real-time context from Primo's unified IT operations platform without switching between consoles or running manual queries. Works with Claude and other MCP-compatible clients. Launched July 10, 2026.

Generally available MCP server from SnapLogic that exposes existing SnapLogic integration pipelines, API Management services, and OpenAPI specifications as MCP tools consumable by any MCP-compatible AI agent. Distinct from the previously released MCP Builder (which creates MCP servers) and SnapCode (which targets coding agents), this is the runtime server that AI assistants connect to for executing enterprise integrations. Includes full governance controls and audit logging. Made generally available July 7, 2026 as part of SnapLogic's July product release.

MCP capabilities from BMC that securely connect AI agents to enterprise workflows and operational data across BMC Helix and mainframe environments. Enables organizations to move from insight to action with visibility, control, and resilience by letting AI agents query and interact with ITSM workflows, operational data, and mainframe operations through governed interfaces. Part of BMC's broader AI agent governance strategy that includes runtime protection and policy enforcement for agent interactions with enterprise systems. Announced July 15, 2026.

Native MCP server from SensorHubb that makes the unified sensor intelligence platform accessible to AI agents. First platform in the sensor intelligence category with native AI-agent access. Teams connect Claude, ChatGPT, Microsoft Copilot, Gemini, or any MCP-compatible agent and query sensor data in plain language — asking which coolers are trending toward failure or which sites logged compliance excursions. Strict safety constraints govern every agent action with confirmation on changes and blocking of destructive operations. Part of SensorHubb 3.0 which also adds predictive equipment health scoring, custom reporting, and building automation connectivity across eight industry protocols including BACnet/IP and Modbus TCP. Serves healthcare, food service, hospitality, education, and life sciences. Launched July 14, 2026.

MCP server from ManageEngine that connects AI assistants to the OpManager Nexus hybrid observability platform for live infrastructure monitoring and AIOps. Once connected to an MCP client like Cursor, VS Code, Claude Desktop, or Windsurf, AI assistants can pull live metrics, find anomalies, and correlate observability data with other tools in the stack including CI/CD pipelines, helpdesk tools, and version control. Ships as a Docker image for on-premises deployments or hosted inside Zoho's private datacenters for cloud customers. Part of OpManager Nexus's broader AIOps capabilities including Zia autonomous agents, smart event correlation, and GenAI integrations with OpenAI, DeepSeek, and Ollama. Announced July 16, 2026.

MCP server that acts as the FinOps cost brain for AI agents in Claude and Cursor, providing comprehensive insights into AWS, Azure, GCP, Kubernetes, and over 15 SaaS and AI provider expenses. Delivers real cost numbers based on actual negotiated rates, identifies genuine savings opportunities, and proposes fixes as actionable pull requests. Runs locally so billing data never leaves the user's machine. Read-only by default with pre-action cost controls that let agents check financial impact before making cost-affecting infrastructure changes. Covers right-sizing cloud resources, analyzing unexpected cost spikes, and implementing cost gates for AI-driven operations.

Open-source MCP server from Zscaler that connects AI agents with the Zscaler Zero Trust Exchange platform for unified security automation. Acts as a secure bridge between AI assistants and Zscaler's APIs, enabling policy queries, configuration management, and multi-step security workflows through natural language. Supports managing several Zscaler products including ZIA, ZPA, and ZDX from Claude, Cursor, or any MCP-compatible client. Designed for security teams who want to automate zero trust operations, investigate incidents, and manage policies using AI agents while maintaining full audit trails.

MCP server from ShareGate that connects the ShareGate Protect Microsoft 365 governance platform to Claude, ChatGPT, and Microsoft Copilot. Enables IT administrators to query governance data including sharing patterns, guest access, workspace activity, sensitivity labels, ownership history, and license usage through natural-language prompts. Tracks trends over time so administrators can see how governance posture is changing. Designed to work alongside other MCP servers like Microsoft Work IQ for richer cross-platform answers. Write actions for remediation planned for July 2026. Included at no extra cost with ShareGate Protect. Launched June 22, 2026.

Enterprise MCP Registry from Workato that gives organizations a governed system of record for discovering, securing, and governing AI tool capabilities at scale. Unlike Workato's existing MCP server for automation execution, the Enterprise MCP Registry is a management layer that catalogs all MCP servers and tools across the organization, enforces access policies, tracks usage, and provides a single control plane for IT and security teams to manage which AI agents can call which enterprise tools. Announced July 16, 2026.

MCP server capabilities within SAP's Business AI platform, part of the Q2 2026 product release announced at SAP Sapphire. Enables AI agents and assistants to connect to SAP enterprise systems through the Model Context Protocol, providing governed access to business data across ERP, supply chain, HR, and finance modules. Designed for enterprise-scale SAP customers who want to connect external AI tools like Claude, ChatGPT, and Microsoft Copilot to their SAP data without building custom integrations. July 2026.

MCP server from Axonius, the asset intelligence platform for unified security operations and exposure management. Connects AI assistants and agents to the Axonius Asset Cloud, translating natural language questions into Axonius Query Language (AQL) and returning live asset data. Gives security and IT teams a governed source of asset truth inside the AI tools they already use, including Claude, ChatGPT and Copilot. Covers devices, users, software, cloud resources, vulnerabilities and exposures across the entire environment. Announced July 21, 2026 alongside the Axonius AI Agent and Docs for AI.

MCP server from Redgate Software for governed, agentic database change management using Flyway Enterprise. Gives AI agents the ability to manage database schema changes with full capture, validation and audit trails. Every AI-driven change is recorded and auditable, addressing governance requirements for teams scaling agentic workflows. Includes advanced Databricks capabilities with Snowflake, operational dashboards and security dashboard views on the roadmap. Part of the broader Redgate AI capabilities platform. Built on Redgate's 25 years of database tooling expertise. Announced July 22, 2026.

MCP server from JAMS, the enterprise job scheduling platform, connecting AI coding assistants to JAMS workload automation capabilities. Built on the open Model Context Protocol standard, it brings JAMS into the AI tools engineering teams already use including Cursor, VS Code with Copilot, Claude Code, Claude Desktop and Codex. Enables developers to manage, monitor and troubleshoot scheduled jobs and automation workflows through natural language. Complements the JAX AI assistant for enterprise job scheduling. Announced July 24, 2026.

MCP server from NowSecure, the mobile application risk management platform, exposing deep mobile app security telemetry to AI assistants and agentic DevSecOps workflows. Gives security teams programmatic access to binary analysis findings, runtime artifacts, vulnerability data, remediation guidance, regulatory compliance mappings and the underlying evidence supporting each risk decision. Enables AI agents to query the NowSecure mobile risk knowledge graph through Claude, ChatGPT or any MCP-compatible client. Part of a broader AI-native update that also includes AI Chat and AI-specific vulnerability detection for mobile apps embedding LLMs. Announced July 23, 2026.

MCP server from Socure, the identity verification and fraud prevention platform, exposing the RiskOS identity risk engine to AI assistants and agentic workflows. Provides programmatic access to identity verification scores, fraud signals, KYC compliance data and risk assessments. Supports token-based authentication with MCP Server Keys replacing username-password auth as of July 21, 2026. Designed for compliance, fraud and operations teams at banks, fintechs and marketplaces.

Open-source MCP server for Vaultwarden and Bitwarden password vaults, built by icoretech. Enables AI assistants and coding agents to securely retrieve credentials, passwords and secure notes from self-hosted Vaultwarden instances or Bitwarden-compatible vaults. Supports both stdio transport for local MCP clients and HTTP SSE for shared service deployments. Installs via npx with environment variables for vault connection. Available on npm as @icoretech/warden-mcp.

High-performance MCP server from Zavora AI for computer use automation, enabling AI agents to interact with desktop applications, browsers and operating system interfaces. Provides tools for screen capture, mouse control, keyboard input and application interaction through the Model Context Protocol. Designed for AI agents that need to perform tasks across desktop software that lacks API access. Open source on GitHub.

Official MCP server from HackerOne, the bug bounty and vulnerability coordination platform, now generally available as of July 2026. Platform-hosted server that enables security teams and engineers to access HackerOne reports, programs, assets, bounties and remediation plans through Claude and other AI assistants. Supports direct authentication with HackerOne accounts. Available on Enterprise, Professional, Legacy and CTEM packaging tiers. Announced July 2026.

Unified MCP server from Klavis AI that enables AI agents to handle thousands of tools reliably at any scale. Acts as a single MCP gateway that aggregates and orchestrates multiple tool integrations, eliminating the need to configure separate MCP servers for each service. Features progressive tool loading, automatic retry and failover, and centralized authentication management. Backed by Y Combinator. Launched on Product Hunt. Designed for teams building agentic AI applications that need many integrations without managing individual MCP servers.

MCP server observability platform from Spanly that provides full visibility into how AI agents interact with your MCP servers. Tracks error rates, session traces, tool call patterns and usage analytics for MCP-enabled products. Designed for teams shipping MCP servers as part of their product, helping them understand agent behavior, debug failures and optimize tool performance. Launched on Product Hunt.

Open-source MCP server from Cyware that brings natural language interfaces to cyber threat intelligence workflows. Integrates with Cyware Intel Exchange and Cyware Orchestrate, enabling SOC analysts to query, summarize and act on threat intelligence data using conversational AI. Supports indicator lookups, CQL search translations, bulk tagging, enrichment metadata, playbook triggering and policy-based automation. Features comprehensive logging, auditability and explainability. Available on GitHub under open-source license.

MCP server from Syncro, the all-in-one PSA and RMM platform for managed service providers. Connects AI assistants to IT service management data including tickets, assets, endpoints and billing. Enables MSPs to query device status, manage tickets and automate IT operations through AI-powered workflows. Part of Syncro's July 2026 release alongside CIS M365 baselines and end user portal SSO. Designed for IT professionals managing multi-client environments who want conversational access to their service management data.

MCP governance and security layer from Snowflake, announced at Black Hat 2026 on July 29. Part of Snowflake Cortex AI Gateway, it provides enterprise-grade controls for AI agent interactions with MCP servers including agent identity management, tool-level access policies, request auditing and data governance. Enables security teams to define which agents can access which MCP tools, enforce data classification policies and maintain audit trails across agentic workflows. Designed for enterprises running AI agents at scale that need centralized MCP governance without modifying individual server implementations.

MCP server from Paragon (useparagon.com) that provides access to prebuilt actions for over 130 SaaS integrations through a single MCP endpoint. Uses SSE transport to support multi-tenant MCP clients. Listed on the Anthropic MCP registry. Supports managed OAuth per end user and per-tenant isolation. Enables AI agents to execute actions across CRM, project management, communication and other enterprise tools through one standardized interface. Backed by SOC 2 compliance. Open source on GitHub under useparagon/paragon-mcp.

MCP server from HashiCorp for Consul, the service mesh and service discovery platform. Enables AI agents to query service catalogs, check health statuses, manage key-value stores and interact with Consul's service networking features through MCP. Note: versions 0.1.0 through 0.1.3 were subject to security advisory HCSEC-2026-24 addressing multiple vulnerabilities, with fixes available in 0.1.4+. Works with Claude Desktop, Claude Code and other MCP-compatible clients.

Security platform from Mondoo that discovers, assesses risk and enforces policy controls for enterprise AI agents and MCP servers. Scans for shadow AI deployments including unauthorized agents, exposed MCP servers and leaked credentials across the organization. Provides vulnerability management, risk scoring and remediation guidance for AI infrastructure. Integrates with existing security workflows. Part of Mondoo's expansion into AI agent governance and security. Announced July 29, 2026.

Open-source MCP server from Tenzir, the security data engineering platform, that uses AI to automatically generate valid Tenzir Query Language (TQL) pipelines for OCSF schema mappings. Takes raw security log samples and returns validated TQL pipelines that map them to the Open Cybersecurity Schema Framework. Built with OCSF at its core. Runs locally for fast processing. Open source on GitHub under tenzir/mcp.

MCP governance and verification platform from JetStream Security that helps enterprises safely adopt third-party MCP servers without unmanaged AI supply chain risk. Combines the JetStream Verified MCP catalog with the JetStream AI Hub to give security, risk and assurance teams one platform to verify server images, control agent permissions and govern runtime activity. Each verified image is produced through proprietary semantic scanning that analyzes data flows and code paths to identify risks such as credential leakage, unicode smuggling and obfuscated malware patterns. FedRAMP Class D High certified. Launched July 13, 2026.

MCP server from JetHost, the Bulgarian hosting provider, that lets customers manage hosting accounts through AI assistants. Connects hosting infrastructure directly to Claude, ChatGPT, Cursor, Lovable and other MCP-compatible tools through a unified MCP framework. Users can check domain renewals, deploy GitHub projects, spin up email accounts, monitor server resource usage and review support tickets through natural language. First Bulgarian hosting provider to offer MCP-based infrastructure management. Launched July 2026.

MCP server from HOSTAFRICA, the first hosting provider on the African continent to support MCP, giving developers direct control of VPS infrastructure from inside Claude, ChatGPT and other MCP-compatible AI tools. Covers the full VPS lifecycle including power management with start, stop, reboot and scheduled actions, backup creation and scheduling, firewall and port configuration, and threshold-based monitoring alerts. VPS instances run in tier three data centres in Johannesburg with nodes in Nairobi, Lagos and Accra. POPIA-compliant with local data residency for financial services, healthcare and government workloads. Launched July 2026.

MCP server from Tenable that exposes 90 structured tools from Tenable's Exposure Data Fabric to any MCP-compatible AI client. Provides a single Tenable-hosted HTTPS endpoint compatible with Claude Desktop, Claude Code, Cursor and any client supporting MCP over HTTP. Enables security teams to build custom agents for unique security workflows while keeping data governed through Tenable's Exposure Data Fabric. Organizations bring their own LLM while Tenable provides ground truth data. Also launched the CyberAgents Exchange, the first open-source cybersecurity-native registry for AI agents, skills, MCP servers and multi-agent playbooks. Announced August 4, 2026 at Black Hat USA.

MCP server from Tanium that exposes governed endpoint data and actions from the Tanium Autonomous IT Platform to MCP-compatible AI clients including Claude, ChatGPT and Microsoft Security Copilot. Security and IT teams can query real-time endpoint intelligence, investigate assets and take remediation actions without leaving their AI workflow. All interactions remain subject to access controls defined within Tanium Atlas. Part of Tanium's broader autonomous security initiative spanning agentic AI, exposure management and security operations. Announced at Black Hat USA August 3, 2026.

MCP server from Intel 471 that connects the Verity471 cyber intelligence platform to AI environments including Claude, ChatGPT and custom agent harnesses. MCP471 exposes underground intelligence across three modes: Intelligence for adversary HUMINT and finished intelligence, Exposure for compromised credentials and vulnerability data, and Hunt for real-time malware indicators. AI agents can resolve actor aliases, pivot across sources and retrieve decision-grade threat intelligence. Pairs with Agent471, a native AI agent inside Verity471 that reasons across intelligence like a human analyst. Announced July 28, 2026.

MCP server from IDrive that enables AI assistants, coding agents and MCP-compatible applications to securely access and interact with data stored in IDrive e2 S3-compatible cloud object storage buckets using natural language commands. Supports OAuth authentication eliminating the need to expose long-term access keys. Provides granular access controls for organizations to determine what AI applications can access. Compatible with Claude, Cursor, ChatGPT, Grok, Codex and other MCP clients. Announced July 30, 2026.

MCP server from Tidal Cyber that extends the Threat-Led Defense platform into AI assistants and agentic workflows. Provides a secure governed connectivity layer between threat intelligence and approved AI tools. Security teams can explore defensive coverage, assess organizational readiness, prioritize engineering work, evaluate threat exposure and generate leadership-ready summaries using trusted threat intelligence. Supports Claude, ChatGPT, Microsoft Copilot and custom agents. Announced August 4, 2026.

MCP integration from Sedai that extends cloud and AI infrastructure optimization capabilities to external AI assistants and agentic workflows. Pairs with Sed, the conversational AI assistant built into the Sedai platform. Users can query cloud costs, identify savings opportunities, review optimization actions and take infrastructure actions through Claude Code, Microsoft Copilot or custom internal agents. Same capabilities available conversationally inside Sedai are available via MCP. Announced August 6, 2026.

Open-source cybersecurity-native registry from Tenable for AI agents, skills, MCP servers and multi-agent playbooks. The only purpose-built security agent exchange in the market. Launches with over 50 AI components under open-source licenses including agents, skill files and MCP servers for vulnerability management, threat intelligence and security operations. Includes contributions from SentinelOne, Recorded Future, Wiz and others. Hosted at exchange.tenable.com. Launched at Black Hat USA 2026 with AWS and Anthropic support. Announced August 4, 2026.

MCP Gateway from Rubrik that acts as a central security checkpoint for all API and MCP resources within an organization. Part of the Rubrik Agent Cloud platform unveiled at Black Hat USA 2026. Every MCP tool call must pass through three checks before execution: SAGE semantic assessment, runtime security policy verification and agent session authentication with short-lived tokens. Integrates with Okta and Microsoft Entra ID to extend existing enterprise identities to autonomous machine agents. Supports On-Behalf-Of federation to restrict access to specific MCP servers and tools on a per-user and per-group basis. Announced August 4, 2026.

MCP integration from Menlo Security that connects agentic systems to secured browser sessions running in disposable containers within the Menlo Cloud. Provides air-gapped security, filtering, auditing and containment for AI agents that access the web. Neutralizes prompt injection by stripping malicious instructions from HTML before they reach an agent. Prevents data exfiltration by masking PII and sensitive IP at the source. Bridges the API Gap by connecting agents to legacy ERP and CRM systems using a managed abstraction layer. Secures browser assistants like Microsoft Copilot, coding agents like Claude Code and autonomous agents. Expanded at Black Hat USA August 5, 2026.

MCP server from Cyera that connects data and AI security insights from the Cyera platform directly into AI assistants and agents including Microsoft Copilot, Claude Code and Cursor. Security teams can accelerate investigations and threat hunting through plain language queries against Cyera DataPort. Enterprises can build their own security agents on top of the platform to address data security risks at scale. Part of the Cyera AI Security Platform alongside Browser Shield and Data Lineage. Agent Guardian expansion at Black Hat USA 2026 adds discovery of shadow agents and MCP activity across endpoints, SaaS and cloud. Released March 2026.

MCP server from Legit Security that delivers AI-native application security intelligence directly into AI code assistants. Connects seamlessly to Cursor, Copilot, Claude Code and Windsurf. Provides real-time security insights and remediation in natural language by connecting SAST and SCA data into a single context-aware intelligence layer that developers can query conversationally. Detects vulnerabilities and policy violations as AI assistants generate code, preventing issues before committing. Legit was named a 2026 Leader in ASPM. Released April 2026.

Centralized MCP gateway from Natoma that acts as the governance and security layer controlling how AI agents use the Model Context Protocol to connect to enterprise systems. Enforces identity, policy and audit at the tool-call level between AI applications and MCP servers. Provides fine-grained authorization with OAuth 2.1, SSO and SAML and SCIM integration, and dynamic policy enforcement. Automatically detects unmanaged MCP servers and rogue agents across the environment. Includes comprehensive logging, SIEM integration and real-time policy monitoring. Acquired by Snowflake in May 2026 and now powers the Cortex AI Gateway. Also available as a standalone product.

MCP server from Bedrock Data that gives enterprise AI workflows direct access to data risk context from the Bedrock Metadata Lake. Part of the ArgusAI platform which governs the enterprise AI risk surface across three components: AI agents enterprises deploy, MCP servers and connectors that broker access to enterprise systems, and the sensitive data those systems can retrieve. Autonomously discovers, classifies and contextualizes sensitive data across cloud, IaaS, PaaS, SaaS and AI environments at petabyte scale. Expanded at Black Hat USA 2026 with Agent DLP for runtime data loss prevention for AI agents. Showcased IANS MCP Intelligence demo at Black Hat.

Open-source MCP proxy from Promptfoo that sits between AI applications and MCP servers to provide enterprise-grade security, monitoring and access control. Whitelists approved MCP servers, grants granular permissions and monitors for PII and sensitive data exposure. Provides detailed logging and real-time alerts for unauthorized access attempts and policy violations. Includes red-teaming capabilities for testing MCP server security with an evil MCP server for adversarial testing. Part of the Promptfoo open-source LLM evaluation and red-teaming platform with over 22,000 GitHub stars. Demonstrated live at Black Hat USA 2026 at the OpenAI booth. Available on GitHub.

Deception security platform from Acalvio Technologies that deploys decoy MCP servers, decoy RAG systems and decoy AI agents alongside honeytokens and deceptive assets embedded into AI agent workflows. Extends the ShadowPlex platform to detect and disrupt compromised AI agents before they access enterprise assets. If an AI agent goes rogue or is manipulated, deceptive assets detect the misalignment, feed the attacker fabricated data and alert the SOC before real enterprise systems are compromised. Provides real-time detection of malicious manipulation, jailbreak behavior and prompt injection attempts targeting agentic AI environments. Demonstrated at Black Hat USA 2026. Announced August 3, 2026.

Agent governance and risk management platform from KnowBe4 that provides real-time visibility into autonomous AI agent activity with automated threat detection and monitoring of agent execution. Supports Anthropic Claude environments with native support for Microsoft Copilot. Monitors actions, outputs and tool usage for agents that automate tasks, analyze data and connect to third-party applications without modifying the underlying model. Leverages 15 years of behavioral data to combat advanced threats including social engineering, prompt injection and shadow AI. Announced at Black Hat USA 2026 on August 3, 2026.

Runtime enforcement platform from Sweet Security that extends autonomous protection from cloud workloads to the AI agents acting alongside them. Blocks rogue agent behavior in real time at the runtime level. Terminates unauthorized tool calls and sessions at runtime, stops secrets, PII and sensitive data from leaving through an agent, and blocks prompt injections live before they steer an agent off course. Unlike monitoring-only approaches that report bad behavior after the fact, Sweet Security blocks threats in runtime the moment they occur. Demonstrated live at Black Hat USA 2026 at Booth 5721. Announced July 29, 2026.

AI agent governance capability from Zero Networks that enforces the OWASP Least Agency principle at the network level for enterprise AI deployments. Allows enterprises to limit what AI agents can access, what actions they can perform and when human approval is required. Leverages identity-based microsegmentation, policy automation and just-in-time multi-factor authentication to control agent behavior. Research from the Zero Networks 2026 Lateral Movement Exposure Report reveals nearly 80 percent of enterprises have deployed internal AI agents yet two-thirds lack governance policies. Demonstrated at Black Hat USA 2026 at Booth 1852. Announced August 3, 2026.

Open-source MCP server for Nutanix Cloud Platform that enables AI assistants to interact with infrastructure through the Nutanix Prism V4 API. Exposes over 1,000 V4 API operations and provides AI tools with API schemas, code samples and permission information. All AI agent activity routes through the Prism V4 API Gateway so every AI-initiated action inherits the same RBAC policies and audit trails that govern human operators. Supports GitHub Copilot, Claude Code and Cursor. Requires Python 3.11 or later and connects to Prism Central. Released as v0.8 Technical Preview. Announced August 10, 2026.

Revamped MCP server from Intezer that connects AI agents to the Intezer security investigations platform for SOC operations. Sits between detection tools and AI agents, giving Claude, Codex and Cursor access to case histories, workflows, triage logic and prior investigation records. Designed as an operating layer for security operations that accelerates SOC tasks by 10x. Intezer AI SOC investigates every alert at 98 percent accuracy in under two minutes with less than 2 percent escalated to human review. Announced June 18, 2026.

Open-source MCP server from Nutanix that enables AI agents and developer tools to interact with Nutanix Cloud Platform through the Prism v4 API. Allows customers to automate daily cloud operations using natural language through AI assistants like Claude and ChatGPT. Supports VM lifecycle management, cluster monitoring, storage operations and infrastructure provisioning with built-in security and governance. Released as v0.8 technical preview. Written in Python. Announced August 10, 2026.

Beta MCP server from Cisco that provides read-only access to Meraki network data for AI agents and applications. Enables AI assistants to query network device status, organization details, client information and network health directly through the Model Context Protocol. Designed for network administrators and IT teams building AI-powered networking experiences. Supports read-only operations with OAuth-based authentication. Part of Cisco's broader agentic networking strategy. Announced August 11, 2026.

Beta MCP server from Cisco that provides read-only access to Catalyst Center network management data for AI agents and applications. Enables AI assistants to query campus and branch network infrastructure, device inventory, network health scores and issue summaries through the Model Context Protocol. Designed for enterprise network teams managing Catalyst-based infrastructure with AI-powered operations. Supports read-only operations with OAuth-based authentication. Part of Cisco's agentic networking initiative alongside the Meraki MCP server. Announced August 11, 2026.

Early-access MCP server from PDQ that connects supported AI tools including Claude, Microsoft Copilot and VS Code to PDQ endpoint-management capabilities. Lets IT teams investigate endpoint health, answer operational questions, generate reports, trigger software deployments and build automations while preserving PDQ administrator-defined permissions, OAuth authentication and audit logging. Announced August 12, 2026.

MCP server for flespi, the telematics and IoT platform, with Realm-based OAuth support for connecting Claude and ChatGPT without manually sharing a dedicated flespi token. Exposes account, device and platform operations through realm public MCP endpoints, with authorization handled through flespi Realms and ACL access such as POST ai/mcp. OAuth connection guide published August 14, 2026.

TypeScript MCP server for KubeAid-managed Kubernetes clusters with first-class ArgoCD application status, workload diagnostics and GitOps-aware operations over stdio. Read tools cover contexts, apps, namespaces, nodes, pods, deployments, events and logs; write paths are explicit and designed around KubeAid dry-run and protected-context controls. Public GitHub repository created August 19, 2026.

Rust policy-enforcing gateway for the Model Context Protocol that aggregates upstream MCP servers, applies tool-call policies, scans results for prompt injection or leaked credentials and writes JSONL audit records. Clients point at portcullis instead of raw servers, enabling default-deny authorization, output sanitization and explainable policy checks before calls reach upstream tools. Public GitHub repository created August 19, 2026.

Hosted Streamable HTTP MCP server for secureFlows that exposes OpenAPI operations tagged ai-safe or ai-optional and forwards calls to the secureFlows backend with caller-provided connection and auth context. It is stateless, keeps tenant configuration out of server storage, and is served at the product /mcp endpoint. Public GitHub mirror created August 20, 2026.

Drop-in MCP server and execution-metrics recorder for Bedrock-based agents and Claude Code. It records per-session cost, token and tool metrics, then renders a Context Window Explorer over a real MCP handshake so teams can inspect visible context and invisible overhead. Public GitHub repository created August 18, 2026.

MCP server for Jira Service Management Operations, focused on the alerts and on-call surface that replaced Opsgenie. It can search alerts, read notes and activity timelines, acknowledge or close alerts, add responders and inspect current or upcoming on-call schedules. Public GitHub repository created August 21, 2026.