Packages

Provides an MCP (Model Context Protocol) server that wraps the dufs API to perform file operations (upload, download, list/search, delete, create directory, move/rename, compute SHA256, and download a folder as a ZIP) using either stdio (MCP-native) or HTTP/SSE endpoints. Supports Basic Auth and configuration via environment variables.

Provides an MCP server with tools to estimate Google Cloud (GCP) running costs by querying Google Cloud Billing Catalog API for SKUs/pricing and (separately) fetching free-tier information from GCP documentation, then calculating estimated monthly costs with automatic free-tier deduction.

A Go SDK to build MCP (Model Context Protocol) servers, letting developers expose MCP tools, resources, and prompts over multiple transports (stdio and HTTP with SSE), while handling MCP protocol message routing and lifecycle in accordance with the MCP specification.

Provides a Go-based MCP (Model Context Protocol) server that exposes markdown files from a local filesystem (via fs.FS) as MCP resources (file:// URIs) and tools to list and read markdown files. Supports YAML (---) and TOML (+++) frontmatter parsing and returns frontmatter metadata in tool outputs/resource descriptions.

mcp-batchit is an MCP “aggregator” server that exposes a single tool, `batch_execute`, to batch multiple downstream MCP tool calls into one JSON-RPC request. It spawns/connects to a target MCP server (e.g., filesystem) and runs sub-operations in parallel up to `maxConcurrent`, with per-op `timeoutMs` and optional `stopOnError` behavior, returning a consolidated result array.

An MCP server that gives AI agents full GUI control over remote macOS machines via Apple Screen Sharing, enabling screenshot capture, keyboard/mouse input, application launching, and drag-and-drop — all without installing software on the target Mac.

Provides an MCP server that helps orchestrate sequential problem-solving by analyzing available MCP tools and returning confidence-scored recommendations (with rationale, priority, and suggested inputs) for each thought/step, including step tracking, branching/revision support, and built-in memory/history limits.

CLI tool that connects to MCP servers via supported transports (command/stdio, streamable HTTP, and deprecated SSE), extracts server capabilities/tools/resources/prompts (optionally calling tools to include results), and outputs documentation in multiple formats (Markdown/JSON/HTML/PDF/Hugo) with customization via config files and templates.

Provides an MCP server (TypeScript) and a Revit add-in (C#) that bridge MCP tool calls from AI clients into Revit actions (read, create, modify, delete elements; also supports exporting/storing some project data).

A Spring Boot sample MCP server that exposes weather-related tools (forecast by lat/long and alerts by US state) via the Spring AI MCP Server WebMVC transport. It can run over HTTP SSE (default) or over STDIO for MCP clients.

Provides a Model Context Protocol (MCP) server that connects to a PostgreSQL database and exposes MCP tools/resources for listing tables, retrieving table schema/sample data, and executing SQL queries through a controlled interface.

pymcp-sse is an asynchronous Python library to build Model Context Protocol (MCP) servers and clients that communicate over HTTP/SSE. It provides base classes for server/client, tool registration and discovery, server push/notification support with keep-alives, and a multi-server client abstraction. It also includes an LLM client abstraction for integration with LLM providers (e.g., an Anthropic Claude example).

Studio (Deco CMS) is an open-source control plane for AI agents. It lets you hire/compose agents, connect and govern external tools via MCP (including “Virtual MCPs”), organize agents and tool connections into projects, and provides observability for tokens/costs/latency/errors. It can run locally with embedded PostgreSQL or via a cloud/team mode.

Provides a standards-compliant MCP Registry Server with a central registry catalog and an Admin API for managing sources/registries/entries. It supports multiple backend data sources (API/Git/File/Managed/Kubernetes) and optional PostgreSQL storage, with OAuth/OIDC-based enterprise auth support.

Provides an MCP server that exposes Umami Analytics data (websites, stats, pageviews, metrics, and active visitors) as MCP tools for MCP clients like Claude Desktop, VS Code/Copilot, Cursor, and others. Supports local stdio transport or an HTTP Streamable endpoint (/mcp) with session handling.

Unity editor tooling that exposes MCP “tools” to create and modify Unity AnimationClips and AnimatorControllers (e.g., create/get/modify .anim and .controller assets) via natural-language commands, built on top of the “AI Game Developer” Unity-MCP platform.

Provides a Python MCP server exposing tools to query a Qdrant vector database using text queries. It can auto-generate embeddings via an external Ollama server and supports advanced retrieval techniques such as query expansion (with RRF merging), HyDE, and LLM-based reranking.

B站（Bilibili）MCP Server that exposes ~27 tools for authentication, search/collection (videos, comments, subtitles, danmaku), content posting (dynamic, video upload, opus), analytics/rankings, and messaging/interaction via MCP clients (e.g., OpenClaw, Claude Code, Cursor, Cline).

Chessmata is a multiplayer chess platform with a browser 3D frontend, a WebSocket-based real-time game/matchmaking system, and programmatic access via both a REST API and an MCP server for AI agent tool-calling workflows. It includes a Python CLI with an MCP server and a UCI adapter to integrate existing UCI chess engines.

Provides an MCP (Model Context Protocol) server exposing tools/resources that query a hybrid GraphRAG datastore: Neo4j for graph-based context expansion and Qdrant for vector/semantic search, with hybrid retrieval combining both.