MCPC

v1.0.0Developer Toolsstable

An extension to MCP (Model-Context-Protocol) that enables two-way asynchronous communication between LLMs and tools through the already existing MCP transport - no additional transport layer needed.

ai-agentsbashclaudeclicode-mode
Share:
642
Stars
0
Downloads
0
Weekly
0/5

What is MCPC?

MCPC is a Model Context Protocol (MCP) server that allows AI assistants like Claude, Cursor, and VS Code to extension to mcp (model-context-protocol) that enables two-way asynchronous communication between llms and tools through the already existing mcp transport - no additional transport layer needed.

An extension to MCP (Model-Context-Protocol) that enables two-way asynchronous communication between LLMs and tools through the already existing MCP transport - no additional transport layer needed.

This server falls under the Developer Tools category on MCPgee, the world's largest MCP server directory with 33,000+ servers.

Features

  • An extension to MCP (Model-Context-Protocol) that enables tw

Use Cases

Enable two-way asynchronous communication between LLMs and tools via MCP. Extend MCP with bidirectional streaming capabilities. Build responsive AI applications with synchronous tool interaction.
apify

Maintainer

LicenseApache-2.0
Languagetypescript
Versionv1.0.0
UpdatedMay 22, 2026
Statushealthy
Maintenanceactive

Works with

ClaudeOpenAIwindowsmacoslinux

Installation

Manual Installation

npx mcpc

Configuration

Configuration Details

Config File

claude_desktop_config.json

Performance

Response Metrics

Response Time< 200ms
ThroughputMedium

Resource Usage

Memory UsageLow
CPU UsageLow

How to Set Up and Use MCPC

MCPC (MCP Client) is a universal command-line client for the Model Context Protocol developed by Apify. It supports both stdio and Streamable HTTP transports, full OAuth 2.1 authentication, persistent sessions, and two-way asynchronous communication between LLMs and MCP tools — all without requiring any additional transport layer beyond standard MCP. Developers use MCPC to connect to any MCP server from a terminal, script tool calls, discover available skills, and manage long-running agentic sessions across providers.

Prerequisites

  • Node.js 18 or later, or Bun runtime
  • npm or Bun package manager
  • Access credentials (OAuth or token) for any MCP servers you intend to connect to
  • An MCP-compatible server to connect to (local or remote)
1

Install MCPC globally

Install the @apify/mcpc package globally so the mcpc command is available system-wide. Both npm and Bun are supported.

npm install -g @apify/mcpc
# Or with Bun:
bun install -g @apify/mcpc
2

Connect to an MCP server

Use mcpc connect to establish a persistent session to a remote or local MCP server. The @session-name makes the session referenceable in subsequent commands.

mcpc connect mcp.apify.com @apify
3

List available tools on the connected server

Once connected, list all tools exposed by the server to understand what capabilities are available.

mcpc @apify tools-list
4

Call a specific tool

Invoke any tool by name and pass arguments as key:=value pairs. String values use key:value and JSON values use key:=value.

mcpc @apify tools-call search-actors keywords:="web crawler"
5

Configure MCPC environment (optional)

MCPC stores state in ~/.mcpc/ by default. Override this or enable verbose logging with environment variables.

export MCPC_HOME_DIR=/custom/state/dir
export MCPC_VERBOSE=1
6

Add MCPC to your MCP client config for AI access

If you want an AI assistant like Claude to orchestrate MCPC sessions, register it as an MCP server in your client configuration.

{
  "mcpServers": {
    "mcpc": {
      "command": "npx",
      "args": ["mcpc"],
      "env": {
        "MCPC_VERBOSE": "1"
      }
    }
  }
}

MCPC Examples

Client configuration

Register MCPC as an MCP server in claude_desktop_config.json for AI-driven session management.

{
  "mcpServers": {
    "mcpc": {
      "command": "npx",
      "args": ["mcpc"],
      "env": {
        "MCPC_HOME_DIR": "/Users/you/.mcpc",
        "MCPC_VERBOSE": "0"
      }
    }
  }
}

Prompts to try

Commands and prompts that demonstrate MCPC's multi-server, bidirectional communication capabilities.

- "Connect to mcp.apify.com and list all available tools"
- "Search for 'web scraping' actors on Apify using the connected session"
- "List all active MCPC sessions and show which tools each one exposes"
- "Reconnect the @apify session and retry the last failed tool call"
- "Use JSON mode to fetch tool results and pipe them to jq for filtering"

Troubleshooting MCPC

mcpc connect fails with authentication error

Run 'mcpc login <server>' first to complete the OAuth 2.1 flow and store credentials. MCPC supports PKCE and Dynamic Client Registration; credentials are persisted in the OS keychain or ~/.mcpc/credentials.

On a headless Linux server, keychain errors appear at startup

MCPC falls back to file-based credential storage at ~/.mcpc/credentials (mode 0600) when the Secret Service API is unavailable. Ensure the directory exists and has correct permissions.

Tool calls hang or time out

Use MCPC_VERBOSE=1 to inspect the request/response cycle. Check that the MCP server is reachable and that the transport type (stdio vs HTTP) matches the server's configuration.

Frequently Asked Questions about MCPC

What is MCPC?

MCPC is a Model Context Protocol (MCP) server that extension to mcp (model-context-protocol) that enables two-way asynchronous communication between llms and tools through the already existing mcp transport - no additional transport layer needed. It connects AI assistants to external tools and data sources through a standardized interface.

How do I install MCPC?

Follow the installation instructions on the MCPC GitHub repository. Clone the repo, install dependencies, and add the server config to your AI client.

Which AI clients work with MCPC?

MCPC works with all major MCP-compatible AI clients including Claude Desktop, Claude Code, Cursor, VS Code (GitHub Copilot), Windsurf, and Cline.

Is MCPC free to use?

Yes, MCPC is open source and available under the Apache-2.0 license. You can use it freely in both personal and commercial projects.

Browse More Developer Tools MCP Servers

Explore all developer tools servers available in the MCPgee directory. Each server includes setup guides for Claude, Cursor, and VS Code.

Quick Config Preview

{ "mcpServers": { "mcpc": { "command": "npx", "args": ["-y", "mcpc"] } } }

Add this to your claude_desktop_config.json or .cursor/mcp.json

Read the full setup guide →

Ready to use MCPC?

Browse our complete directory of 33,000+ MCP servers, read setup guides for your editor, and start building with the Model Context Protocol.

33,000+ ServersFree & Open SourceStep-by-Step Guides