OPLINK
Getting Started

Introduction

Intro to the OPLINK docs

Welcome to OPLINK, an MCP server for workflows. It sits between your IDE and multiple MCP servers, reads simple YAML workflow definitions, and exposes them as a single, well‑structured tool surface in your editor.

Model Context Protocol (MCP) lets AI agents call tools directly from your editor. But as soon as you add more than one MCP server, things get messy:

  • Each server has its own tools, names, and auth quirks.
  • Prompts and workflows end up copy‑pasted into different projects.
  • There’s no single place to version and share “how we actually debug, triage, or investigate things.”

OPLINK fixes this by acting as the workflow hub:

  • You declare workflows in .mcp-workflows/workflows.yaml.
  • Oplink runs as an MCP server (oplink server --config ./.mcp-workflows).
  • Your IDE/agent sees one MCP endpoint with curated tools (your workflows), not a pile of raw servers.

Key Features

Workflow Orchestration

Combine multiple MCP servers and prompts into powerful, reusable workflows

Custom Commands

Trigger workflows with intuitive commands like "enter debugger mode" or "use thinking mode"

Version Control

Share and version control your workflows through simple YAML files

Smart Strategies

Define custom strategies for using tools sequentially or situationally

Team Onboarding

Easily onboard your team to best practices with versioned workflow configurations

Examples Library

Curated example workflows for common development tasks

Why OPLINK?

OPLINK was born from the realization that “just add more MCP servers” doesn’t scale on its own. The valuable part is how you sequence and combine those tools for real tasks: debugging, incident response, doc lookup, triage, migrations, etc.

By putting workflows in one MCP server, OPLINK helps you:

  • Standardize development practices with named workflows (frontend_debugger, deepwiki_lookup, etc.)
  • Reduce cognitive load when switching between tasks (agents and humans see a small, meaningful surface)
  • Make AI tools more accessible and practical for your team
  • Improve collaboration and knowledge sharing through versioned configs
  • Accelerate development cycles with reusable, reviewed flows

Whether you're a solo developer or part of a large team, OPLINK turns your MCP ecosystem into a coherent, versioned set of workflows instead of a bag of individual servers.

Thanks

Special thanks to:

  • Peter Steinberger for bringing in and maintaining mcporter, which powers Oplink’s external MCP discovery and transport.
  • Ted Werbel with which I've started this project originally in early 2025 as MCPN. It gave a base idea what oplink is today

Installation

Installing OPLINK