Packages

Upsonic is a Python agent framework for building production-oriented LLM agents. It supports multiple model providers (e.g., OpenAI/Anthropic/Azure/Bedrock), tool integration (including MCP and custom tools), OCR/document processing, memory/state management, multi-agent coordination, and a safety engine with policy-based content handling.

A Python Model Context Protocol (MCP) server that provides tool-based access to VMware vSphere (vCenter/ESXi) for infrastructure management. It exposes MCP tools for authentication credential caching, VM lifecycle operations, inventory lookups (hosts/datacenters/datastores), folder organization browsing, and network/VLAN discovery using the vSphere APIs.

Provides an MCP server that lets AI assistants interact with Yandex Tracker: queue and field metadata, searching, and full issue lifecycle (read/create/update/transition/close), with optional Redis caching and OAuth2-based authentication and token refresh. Supports multiple transports including stdio, SSE (deprecated), and HTTP.

Provides an MCP server that connects a user's Zotero library to AI assistants, enabling semantic/keyword search, retrieval of item metadata/full text/attachments and annotations, and write operations (e.g., add items by DOI/URL/file, manage collections/tags/notes). Supports local Zotero access, web API access, and an MCP transport option (stdio/streamable-http/sse). Optional extras add semantic search embeddings, PDF outline/annotation extraction, and Scite citation intelligence.

AgentAnycast is a P2P runtime (SDKs plus a local Go daemon) for connecting AI agents across networks without requiring public IPs. It provides NAT traversal, end-to-end encrypted task/message transport, peer identity (Ed25519 keys mapped to W3C DIDs), and routing by agent skill/capability via a self-hostable relay and optional mDNS discovery on LAN. It also exposes an MCP server for tool/skill interoperability and offers an HTTP bridge for HTTP-based A2A agents.

Cognio is an MCP server and local REST API for storing persistent semantic memories in a SQLite database, enabling semantic (and optional hybrid) search across conversations/projects. It also exposes MCP tools for memory CRUD, project management, stats, export, and text summarization, plus a local web UI dashboard.

HopperPyMCP is a Python MCP (Model Context Protocol) server intended to be installed as a Hopper disassembler script/plugin. It exposes Hopper analysis capabilities as MCP tools, including binary/document management, segment and address information, regex-based name/string search, disassembly and decompilation, call graph generation, cross-reference/reference analysis, and annotation operations (comments/names/types) on the current Hopper document.

Provides an MCP (Model Context Protocol) server exposing Kubernetes cluster interactions (discovery/listing/details/describe, logs, node/pod metrics, events, and resource create/update/delete), with optional Helm-related tooling and a read-only mode. Can run over stdio, SSE, or streamable-http transports.

meGPT is a Python repository for ingesting an author’s public content (books, blog posts, social archives, YouTube, podcasts, etc.), preprocessing it (downloading/extracting/transcribing/summarizing/deduplicating), and exposing the resulting knowledge base via a Model Context Protocol (MCP) server for AI apps to search and retrieve content in an “author voice”/persona style.

Provides an MCP server that bridges AI assistants to an MS SQL Server database for discovery (schemas/tables), read-only SQL querying, and natural-language-to-SQL, with optional stdio (default) and HTTP/SSE transports.

PlexMCP is an MCP (Model Context Protocol) gateway that orchestrates MCP servers behind a unified API/dashboard. It provides multi-tenant isolation, authentication (API keys and TOTP 2FA), audit logging, and usage analytics/billing, with a Rust/Axum backend and Next.js/TypeScript frontend.

Provides an MCP server that lets an MCP client (e.g., Cursor) control web browsers via Selenium WebDriver, exposing tools for navigation, interaction, waiting, screenshots, and extracting page information like an accessibility snapshot.

Provides a Model Context Protocol (MCP) server that exposes Cloud Foundry operations as 38/associated tools (apps, orgs/spaces, services, routes, network policies, cloning, and target/validation). Intended to be deployed on Cloud Foundry and accessed by an MCP client via the streamable HTTP transport.

Provides an MCP-compatible server to control a Dynatrace UFO lighting device from MCP clients (e.g., Claude Desktop), supporting multiple transports (stdio, HTTP/2+SSE, and legacy SSE). Includes a library of built-in lighting effects and tools for querying/controlling UFO LED state and sending raw commands.

vLLM MCP server that exposes vLLM chat/completions plus model and status/metrics capabilities to MCP-compatible clients. It can also start/stop/manage a vLLM instance in Docker/Podman with platform/GPU-aware container selection and runtime configuration via environment variables and MCP tools.

Provides an MCP server (Python) exposing tools to read, write, create, and update Excel 97-2003 .xls (binary) files, including listing sheet names and retrieving basic spreadsheet metadata.

Provides a Model Context Protocol (MCP) server over stdio (stdin/stdout) that exposes X/Twitter functionality as MCP tools, including basic user lookup, tweet retrieval/search, and (per README) posting with OAuth 1.0a/Bearer-token-style configuration via environment variables.

Provides an MCP server/client in Python that lets AI assistants search, read, create, update, delete, and bulk-manage records in an Odoo ERP instance via MCP tools mapped to Odoo model operations. Configuration is done through environment variables, including an Odoo API key for authorization, and supports MCP stdio or streamable-http transport modes.

MCP server bridging AI clients (Cursor, Claude Desktop) to Agile Luminary project management system. Provides three tools to retrieve project details, current work assignments, related documents, and historical task context from Agile Luminary API.

Provides a Model Context Protocol (MCP) server that exposes Alpha Vantage stock/market data as MCP tools. Can run as an stdio MCP server (default) or as an HTTP “streamable” MCP server, with optional OAuth 2.1 protection for the HTTP mode. Uses an Alpha Vantage API key via environment variable `ALPHAVANTAGE_API_KEY`.