The open registry for MCP servers

A Model Context Protocol server that provides read and write access to local SQLite databases. Enables LLMs to execute SQL queries, inspect schemas, and manage data through standard database operations.

Interact with GitHub APIs through MCP. Manage repositories, issues, pull requests, code search, and file operations — giving autonomous agents full-cycle software development capabilities.

Secure, configurable filesystem access for MCP agents. Supports reading, writing, listing, searching, and moving files — with sandboxed directory restrictions for safe autonomous file operations.

Web and local search capabilities via Brave Search API. Enables agents to perform real-time web searches, news lookups, and local business queries with privacy-first, ad-free results.

Attack surface reconnaissance via Shodan's internet-wide scanning database. Query exposed services, CVEs, open ports, and organization-wide exposure reports — purpose-built for security automation.

Network discovery and port scanning through MCP. Enables security agents to run configurable Nmap scans, parse results, and feed structured host/service data into automated vulnerability workflows.

Read-only PostgreSQL access for LLM agents. Inspect schemas, run analytics queries, and retrieve relational data with safe, connection-pooled access to production or staging databases.

Manage Docker containers, images, volumes, and networks via MCP. Gives DevSecOps agents the ability to spin up containers, run workloads, and manage container lifecycles.

SAST (Static Application Security Testing) via Semgrep rules directly in MCP. Agents can scan codebases for security anti-patterns, secrets, and vulnerabilities as part of automated code review pipelines.

Send messages, read channels, manage workspaces, and search Slack history. Lets orchestration agents deliver alerts, status updates, and incident notifications into team communication channels.

Headless browser automation through MCP. Capture screenshots, fill forms, scrape dynamic pages, and automate web interactions — essential for agents that need to test or monitor web applications.

Retrieval-Augmented Generation (RAG) against Amazon Bedrock Knowledge Bases. Provides agents with semantic search over enterprise document stores, FAQs, and internal wikis stored in AWS.

Structured step-by-step thinking and planning capabilities. Allows agents to break down complex tasks, plan execution paths, evaluate hypotheses, and dynamically adjust reasoning strategies.

Persistent semantic knowledge-graph memory for AI agents. Enables LLMs to construct, read, write, and query fact-relationship networks to maintain long-term memory across chat sessions.

Fetch web page content and convert it into clean, readable Markdown. Optimized for LLMs to read documentation, APIs, and online articles without HTML or script noise.

Interact with Notion workspaces through MCP. Retrieve page contents, query databases, append blocks, and search workspace pages to give agents access to team wikis and document hubs.

Linear project tracker integration. Search issues, create new tasks, update tickets, and inspect backlogs to help autonomous project management agents orchestrate development cycles.

Interact with GitLab API. Manage repositories, issues, merge requests, groups, and build pipelines — giving autonomous agents full-cycle GitLab development capabilities.

Interface with Sentry to retrieve application issues, stack traces, and release logs. Enables automated debugging agents to discover and analyze production errors.

Provides agents with geo-location, search, mapping, and direction calculations using the Google Maps platform APIs. Ideal for travel and logistics agent routing.

A local stdio MCP server that lets AI coding assistants execute shell commands on remote computers through an HTTP relay. It supports per-command approval or automatic execution, while the remote machine needs only curl and bash.