MCPSERV.CLUB
endorhq

Endor MCP Server

MCP Server

Instant sandboxed services for AI agents and developers

Stale(55)
24stars
1views
Updated 23 days ago

About

Endor launches private, isolated database and KV store environments instantly on any Node runtime. It supports MCP integration so AI tools can securely start services like MariaDB, Postgres, Redis, and more with minimal setup.

Capabilities

Resources
Access data sources
Tools
Execute functions
Prompts
Pre-built templates
Sampling
AI model interactions

Endor MCP Server Overview

Endor is a lightweight, zero‑install solution that spins up fully isolated database and messaging services on any machine running Node.js. It addresses the common pain point of provisioning temporary, secure environments for AI agents and developers who need instant access to services like MariaDB, PostgreSQL, Redis, or RabbitMQ without the overhead of Docker, cloud accounts, or manual setup. By launching each service in an isolated VM that exposes only the application ports, Endor guarantees that nothing persists after you terminate the session—ideal for privacy‑sensitive workflows and rapid prototyping.

In MCP mode, Endor exposes a rich set of capabilities to AI assistants: launching services on demand, monitoring their health, and providing connection details through a standardized protocol. This means an agent can, for example, start a Valkey instance as part of a data‑processing routine and then immediately query it without any manual configuration. The server’s ability to run a full Alpine Linux environment in experimental networking mode further extends its utility, allowing agents to execute arbitrary commands or install additional tools within a clean sandbox.

Key features include:

  • Instant, private sandboxes for databases, caches, and message brokers.
  • Ephemeral isolation: each service runs in its own VM with only the necessary ports exposed.
  • MCP integration: agents can request, monitor, and terminate services through a standard protocol.
  • Zero‑install CLI: the package installs locally and works with , eliminating global dependencies.
  • Experimental full‑networking mode: a minimal Alpine Linux VM for advanced use cases.

Typical use cases span both AI and developer workflows. A Claude Code assistant can spin up a PostgreSQL instance to test SQL queries, while a VSCode extension can automatically provision a Redis cache for debugging. In continuous‑integration pipelines, an agent could start a RabbitMQ broker, run integration tests, and clean up instantly. The sandboxed nature also makes Endor suitable for security‑critical environments where accidental persistence of data is unacceptable.

Endor’s standout advantage lies in its simplicity and security. By abstracting away the complexity of container orchestration or cloud provisioning, it lets developers focus on building logic rather than managing infrastructure. The MCP interface ensures that any AI tool—whether a code editor, chatbot, or custom workflow engine—can seamlessly request and control services, making Endor a versatile bridge between local development environments and intelligent assistants.