43.008 servidores que conectas por URL mediante streamable HTTP. Nada que instalar ni mantener en ejecución en tu máquina.
Pega su repositorio de GitHub y lo listaremos tras una revisión rápida.
Un MCP server remoto se ejecuta en la infraestructura de otra persona y tu cliente se conecta a él por URL mediante streamable HTTP (los servidores más antiguos usan SSE). El inicio de sesión suele hacerse con OAuth en el navegador en lugar de una clave API en un archivo de configuración.
Los servidores locales se ejecutan como un proceso en tu máquina y pueden acceder a tus archivos y red privada. Los servidores remotos no necesitan un runtime y se actualizan solos, pero tus datos van al proveedor. Elige remoto para productos SaaS y local para archivos, bases de datos en tu red y todo lo que funcione sin conexión.
En Claude Code ejecuta claude mcp add --transport http <name> <url>. En Cursor y Gemini CLI agrega la URL en mcpServers; en VS Code agrégala en servers con "type": "http". Cada página de servidor tiene la configuración exacta.
Lets an AI agent store, recall, and calibrate persistent memories, search indexed documents, and report outcomes.
$claude mcp add mcp -- npx -y memcell
Lets AI agents save, inspect, correct and forget memories across sessions, keeping a Source Receipt for each one.
$claude mcp add --transport http cairn-memory https://cairn.ink/api/mcp
Provides persistent, governed long-term memory and session continuity for AI agents across different tools and sessions.
$claude mcp add memory-os-cli -- npx -y @xmemo/client
Provides an agent-agnostic memory backend that preserves continuity between humans and AI using streamable HTTP transport.
$claude mcp add marsnme -- npx -y @marsnme/mcp-gateway
Provides persistent, scored memory for AI agents by allowing Claude and other clients to read, write, and manage a shared brain.
$claude mcp add --transport http brain https://mcp.brainmemory.ai/mcp
Provides secure durable memory and bounded context for AI agents through local or hosted knowledge management and hybrid search.
$claude mcp add --transport http graymatter https://api-0.valkyrlabs.com/graymatter/mcp
Provides a private memory layer for AI assistants that stores and recalls information locally on your computer.
$claude mcp add modelbrain-mcp -- npx -y modelbrain-mcp
Provides a persistent, cross-vendor semantic memory store that allows AI clients to read and write facts across different LLM platforms.
$claude mcp add --transport http gnosem https://gnosem.dev/mcp
Provides persistent, cross-tool memory for AI agents to save, search, and connect context across different sessions and projects.
$claude mcp add --transport http empirical-memory-plugin https://empirical.gauzza.com/mcp
Provides AI assistants with persistent, searchable, and portable long-term memory across different conversations and clients.
$claude mcp add --transport http amber https://mcp.ambermem.com/
Provides AI clients with brain-inspired long-term memory through hybrid retrieval, evolving user profiles, and multimodal data ingestion.
$claude mcp add --transport http zhiji-mcp https://ai-know.me/mcp
Provides a semantic memory infrastructure for AI agents to save decisions, context, and sources with quality-gated writes and conflict tracking.
$claude mcp add --transport http cortex-connector https://skynetlab-cortex-saas-mcp.cortex-320.workers.dev/mcp
Provides a shared remote memory for AI tools to recall and store information across different client applications.
$claude mcp add --transport http tephra-mcp https://tephra.obsidian-metrics.com/mcp
Provides persistent conversational memory, behavioral code search, and automated bug localization for AI coding agents.
$claude mcp add --transport http ml-memory https://api.slashml.ai/mcp
