MCP Terminal 是一个基于 MCP（Model Context Protocol）的终端控制服务器，专为与大型语言模型（LLM）和 AI 助手集成而设计。它提供了一个标准化的接口，使 AI 可以执行终端命令并获取输出结果。

Hybrid MCP integration blending Vercel AI SDK UI with standardized context enrichment, enabling modular, real-time AI apps with pluggable data sources.

TypeScript-based, memory-backed, multi-provider chat with persistent history across OpenAI, Anthropic, OpenRouter, and Ollama.

A collection of MCP servers enhancing Claude desktop capabilities with screen capture, automation, FastAPI data processing, and DuckDB analytics.

Transforms prompts into CoD/CoT, supports BYOLLM, and enables efficient, flexible reasoning across cloud and local models.

Path of Building MCP

A MCP server to interact with Kalshi prediction markets

OpenAI-powered collection tool that discovers MCP endpoints from URLs or content and submits them to a central directory.

Fast atomistic modeling toolkit offering structure generation, MLIP-based optimization, MD simulations, and artifact delivery for streamlined workflows.

Blazing-fast native Go implementation bridging Colab and local workflows with static tooling and seamless WebSocket communication

Distinctive local tooling for Elegoo Centauri Carbon delivering offline automation, dashboards, and camera streaming through a unified Python library and UI.

Local-first markdown vault with a deterministic SQLite index and active, executable contracts guiding agent-driven workflows.

Persistent terminals with real PTY, incremental per-consumer reads, headless screen rendering, and browser dashboard for multi-agent AI workflows.

Bridge IMAP/SMTP mailboxes to Poke with multi-account support, real-time updates, and secure per-server authentication; emphasizes non-destructive handling and automatic Poke-forwarding.

Unified framework blending Perplexity skill creation with Claude Code orchestration to design modular AI agents and coordinated workflows.

Rust-based AI access gate enforcing SPF governance with deterministic risk scoring and comprehensive tool controls, distinguishing it through real-time gating and robust content inspection.

Code Mode turns Unity Editor into an AI-empowered workspace by exposing scene, asset, inspector, and editor actions as UTCP tools. Tools are discovered automatically, executed on Unity's main thread, and driven by JavaScript definitions to minimize tokens while enabling complex automation across environments.

Issues temporary, service-scoped AWS credentials via STS for AI agents to securely access resources on demand.

Open-source knowledge backend delivering persistent, container-scoped memory with hybrid search, runtime configurability, and multi-provider embeddings for Claude integration.

An MCP server that lives inside Rider and exposes Resharper-powered tools for AI agents that allow working with actual codebase semantics instead of `grep` approximations.

FigMCP

Text-to-video animation via Manim. Text → Code → Video. CLI, agent mode, and MCP server

XRP Ledger MCP Server

deprecated: use the official MCP sdk! (https://github.com/modelcontextprotocol/go-sdk) / ⚡ A type-safe, intuitive Go SDK for building MCP servers with ease and confidence