Integrations

MCP Servers

OPIDE supports the Model Context Protocol. Connect any MCP server and its tools become available to the agent alongside the built-in tools.

What MCP is#

Model Context Protocol is an open standard for connecting AI agents to external tools and data sources. An MCP server exposes a set of tools that any compatible agent can call. OPIDE implements the full MCP specification on the client side.

This means any MCP server built for Claude, Cursor, or any other compatible client works in OPIDE without modification. The ecosystem is shared.

Connecting a stdio server#

Stdio servers run as child processes. OPIDE spawns them, communicates over stdin/stdout, and manages their lifecycle. This is the most common MCP server type.

Add a stdio server in Settings under MCP Servers:

◆Name — a label for the server in the UI
◆Command — the executable to run (e.g. npx or uvx)
◆Args — command arguments (e.g. -y @modelcontextprotocol/server-filesystem)
◆Env — optional environment variables passed to the server process

Example: Filesystem MCP server config

{
  "name": "filesystem",
  "command": "npx",
  "args": ["-y", "@modelcontextprotocol/server-filesystem", "/Users/you/projects"],
  "env": {}
}

Connecting an SSE server#

SSE (Server-Sent Events) servers run remotely and expose an HTTP endpoint. OPIDE connects via long-polling and handles reconnection automatically.

◆Name — label for the server
◆URL — the SSE endpoint (e.g. http://localhost:3001/sse)
◆Headers — optional auth headers

ℹInfo

SSE servers are useful for shared team tooling — a single MCP server running in your infrastructure that all developers connect to from their own OPIDE instance.

Tool discovery#

When OPIDE connects to an MCP server it automatically discovers all available tools, their schemas, and their descriptions. The agent sees these tools alongside the built-in OPIDE tools. No manual configuration of individual tools is needed.

MCP isolation#

External MCP servers run in separate processes isolated from the OPIDE engine. Tool results from MCP servers are validated against the declared schema before being injected into the agent context. A misbehaving or compromised MCP server cannot affect the OPIDE process.

⚠Warning

Only connect MCP servers you trust. Tool results from external servers are presented to the agent as data — a malicious server could attempt to inject instructions via tool results. OPIDE validates schema compliance but cannot validate intent.

← PreviousAST Indexing

Next →Open VSX Extensions