Packages

An example MCP (Model Context Protocol) server/client implementation that demonstrates exposing MCP tools, resources, and prompts, with a simple knowledgebase RAG flow backed by ChromaDB. Includes a CLI client to connect to a local MCP server instance.

SpecLock is an AI constraint/policy enforcement system for coding agents. It defines text and typed constraints (locks), reviews intended changes (including diff-based review), blocks or warns on violations via a hard/enforcement mode, and provides auditing (HMAC chain) plus encryption and RBAC for enterprise use. It also exposes MCP tooling and a REST API for automated review/checking, with a Python SDK and ROS2 integration described.

A locally hosted, read-only MCP server that exposes SurveyMonkey data via CData’s JDBC Driver as MCP tools (get tables, get columns, run SQL SELECT queries). It’s intended for use with MCP-capable desktop clients (e.g., Claude Desktop) and requires local configuration using a .prp file that points to the CData JDBC driver and connection/JDBC URL (including OAuth parameters if used).

Provides a local (stdio) read-only Model Context Protocol (MCP) server that wraps the CData JDBC Driver for Twilio, exposing Twilio data as relational tables/columns and tool-callable operations (list tables/columns, run SELECT queries) for AI clients like Claude Desktop.

Provides an MCP server (Node.js) that exposes Unreal Engine 5 editor operations (levels, actors, Blueprints, Animation Blueprints, assets, viewport capture, and console/script execution) to any MCP-compatible AI client. It bridges MCP tool calls to a separate Unreal HTTP backend (default http://localhost:3000) via an expected REST API.

An MCP server that exposes tools to create, edit, visualize (ASCII/PNG), decode, and play VibeTide 2D platformer levels (including tile-level editing and metadata management) via the Model Context Protocol.

zero-mcp is a lightweight TypeScript toolkit for implementing MCP (Model Context Protocol) servers that expose JSON-RPC tool endpoints over native HTTP. It provides a McpServer abstraction with tool registration (Zod-based schemas + JSON schema generation) and lifecycle hooks, plus CORS controls for browser-based clients.

An MCP (Model Context Protocol) server that lets an MCP-supported IDE connect to Figma and use F2C to parse Figma design nodes/links and generate pixel-perfect HTML/CSS and framework-specific code (e.g., React, CSS Modules, Tailwind). It also supports localization of remote Figma images to local assets.

GenFilesMCP is an MCP server (with an HTTP streamable mode and a stdio/MCPO mode) that generates document files (DOCX, PPTX, XLSX, and Markdown) from user requests and chat context, integrates with an Open Web UI (OWUI) endpoint to upload/attach generated (and reviewed) files, and can also review existing Word documents by extracting structure and adding AI-generated comments for grammar/ideas.

Hoverfly MCP Server is a Spring Boot-based Model Context Protocol (MCP) server that exposes Hoverfly’s mock-management capabilities as MCP tools, enabling AI assistants to start/stop Hoverfly, create/clear/list mocks, fetch status/version, and download simulation data (typically via a Docker deployment with mounted persistence).

MCP Inspector is a developer tool for testing and debugging Model Context Protocol (MCP) servers. It includes a React web UI and a Node.js proxy that bridges to MCP servers over multiple transports (stdio, SSE, streamable-http), exposing a browser-based workflow for configuring, running, and inspecting MCP tools.

log-mcp is a Python-based MCP server that analyzes large log files for AI agents. It provides tools to scan log overviews, search by regex/level/time, extract segments, deduplicate and summarize errors, compute log statistics, compare patterns across files, and classify lines using an optional Rust TF-IDF classifier plus an optional BERT-mini re-scoring stage. It can run with or without the Rust classifier (Python fallback).

MCP Anywhere is a Python-based gateway that discovers and manages MCP servers from GitHub repositories and exposes them via a unified endpoint. It includes web UI/API authentication, tool enablement controls, and Docker-based isolation for running MCP tools, with encrypted secret-file storage for credentials used by MCP servers.

Provides an MCP server template integrated with Mem0 to let AI agents save, retrieve, and semantically search long-term memories backed by Supabase/PostgreSQL (vector storage) and configurable via an LLM provider for embeddings.

nexonco-mcp is an MCP server that exposes a tool (search_clinical_evidence) for querying and returning formatted clinical evidence from the CIViC database via flexible filters such as disease, therapy/drug, molecular profile/variant, phenotype, and evidence metadata.

Photon turns a single TypeScript class ("photon") into multiple runnable interfaces: an auto-generated web UI (Beam) for humans, a CLI tool, and an MCP server for AI agents, deriving tool schemas and validation from TypeScript method signatures, types, and JSDoc/comments.

Provides a read-only Model Context Protocol (MCP) server that exposes Sage 300 data as SQL-like tables/columns via CData JDBC. Intended for LLM clients (e.g., Claude Desktop) to query live Sage 300 data through MCP tools (get_tables, get_columns, run_query).

Provides a sample Spring AI MCP (Model Context Protocol) server implemented with Spring Boot, exposing tool-based CRUD/search operations over an in-memory Person dataset (seeded with dummy/CSV data).

Provides a Model Context Protocol (MCP) server exposing tools for managing Temporal workflow orchestration (start/query/control/history) and schedule management. Supports running against local or remote Temporal instances, with Docker or PyPI-based execution, and configurable connection/TLS/API key settings.

Provides an MCP (Model Context Protocol) server that wraps webhook.site capabilities: creating and managing webhook endpoints (HTTP), and retrieving captured HTTP requests, emails, and DNS lookups. It also exposes tools for waiting/polling for new events and for generating common security test payloads (e.g., SSRF/XSS canary-style artifacts) and exporting captured data.