🍠 YAMCP - A Model Context Workspace Manager
LocalAbout 🍠 YAMCP - A Model Context Workspace Manager
YAMCP is a command-line tool for grouping multiple MCP servers into local workspaces called YAMs, exposed to AI apps through a single gateway. It centralizes server communication logs, which simplifies monitoring and debugging.
Key features
- 01Import, add, list and remove server configurations
- 02Create and manage workspaces by purpose or app
- 03Expose workspace servers through a unified gateway
- 04Store all server communication logs in one place
- 05Optional web interface through YAMCP UI
Use cases
- 01Group servers for coding, design or data work separately
- 02Share one server set across Cursor, Claude and other AI apps
- 03Investigate connection issues across several MCP servers
MCP support
YAMCP acts as both an MCP client and server: it connects to local and remote MCP servers and presents them to AI apps as a single gateway MCP server.
Configuration
Local server config is the same for every client. Copy it into your client’s MCP config file.
Tools
Local servers expose tools only once running on your machine, and the maintainer has not provided a tool list.
Details
From the project README🍠 YAMCP - A Model Context Workspace Manager
YAMCP (YAM-C-P) is a command-line tool for organizing and managing MCP servers as local workspaces. It seamlessly connects to multiple MCP servers, local or remote, grouping them into a unified workspace exposed as Yet Another MCP server (YAM) for AI applications. You can create dedicated workspaces based on specific functionality (e.g., a YAM workspace for coding, design, research, ...) or based on the AI apps that consume servers (e.g., a YAM for Cursor, Claude, Windsurf) or any other combination in between. In addition, it simplifies monitoring and debugging MCP servers by centralizing all server communication logs in a single store, eliminating the need to dig through each AI client app’s logs separately.
Import and Create Workspaces
Connect All Bundled Servers in a Workspace to Your AI Apps with One Config
🚀 Quick Start
# Install YAMCP
npm install -g yamcp # or use npx yamcp
# Import servers (choose one)
yamcp server import [config] # import servers from config file (see src/example-servers.json for format)
yamcp server add # or add manually
# create workspaces (e.g. a yam for coding, design, data, ...)
yamcp yam create
# Run workspace in your AI app
yamcp run <yam-workspace-name>
🔑 Key Concepts
- MCP Servers: Remote or local servers that provide Model Context Protocol services
- Workspaces (YAMs): Collections of MCP servers grouped together to be shared with AI Apps (e.g. a workspace for coding, writing, design, magic making!)
- Gateway: A local MCP server that manages connections to configured MCP servers in a workspace and exposes them through a unified server to AI App's MCP clients
With YAMCP, you can:
- Create workspaces to group MCP servers by AI application (e.g. Cursor, Claude, GitHub Copilot)
- Group servers by workflow purpose (e.g. software development, data science, technical writing)
- Connect AI apps to a single gateway that provides access to all workspace servers
- Manage and monitor multiple MCP server connections through a unified interface
- Track all server communications with detailed logging and debugging capabilities
Top-Level Commands
yamcp [command] [subcommand] [flags]
Available top-level commands:
server- Manage MCP providersyam- Manage workspaces (yams)run- Run the gateway with a workspacelog- View the server log location
🔧 Mcp Server Management Commands
Server Commands
yamcp server add # Add a new MCP server (interactive)
yamcp server list # List all configured servers and their status
yamcp server remove # Remove a server configuration
yamcp server import # Import server configurations from a JSON file
🍠 Yam Workspace Management Commands
Workspace Commands
yamcp yam create # Create a new workspace (interactive)
yamcp yam list # List all workspaces or show specific workspace details
yamcp yam edit # Modify an existing workspace configuration
yamcp yam scan # Scan workspaces
yamcp yam delete # Delete a workspace
Runtime Commands
yamcp run <yam-workspace-name> # Start the gateway with specified workspace
yamcp log # View server communication logs
✅ Command Reference
| Command | Description | Example |
|---|---|---|
server add | Add a new MCP server | yamcp server add |
server list | List configured servers | yamcp server list |
server remove | Remove a server | yamcp server remove [name] |
server import | Import server config | yamcp server import [config] |
yam create | Create workspace | yamcp yam create |
yam list | List workspaces | yamcp yam list |
yam list --name | Show workspace details | yamcp yam list --name my-workspace |
yam edit | Edit workspace | yamcp yam edit |
yam scan | Scan workspace | yamcp yam scan [workspace-name] |
yam delete | Delete workspace | yamcp yam delete [workspace-name] |
run | Start gateway | yamcp run <workspace-name> |
log | View logs | yamcp log |
🖥️ YAMCP UI (created by @eladcandroid)
YAMCP UI provides an intuitive web interface to manage your MCP servers and workspaces through a universal dashboard.
# Run directly with npx (recommended)
npx yamcp-ui
# Or install globally
npm install -g yamcp-ui
yamcp-ui
Refer to the project repo for full documentation: https://github.com/eladcandroid/yamcp-ui.
🏗️ System Architecture
graph TB
CLI[CLI Commands]
GW[McpGateway]
GS[GatewayServer]
GR[GatewayRouter]
LOG[Logger]
STORE[(Store)]
AI_APP[AI App]
%% CLI Command Flow
CLI -->|manages| STORE
CLI -->|runs| GW
%% Gateway Components
GW -->|uses| GS
GW -->|uses| GR
GW -->|logs| LOG
%% Server & Router
GS -->|stdio transport| AI_APP
GR -->|connects to| SERVERX
GR -->|connects to| SERVERY
%% Data Store
STORE -->|loads config| GW
%% External MCP Servers
subgraph "Workspace Servers"
SERVERX["Server x (Stdio)"]
SERVERY["Server y (SSE)"]
end
%% Store Components
subgraph "Configuration Store"
PROVIDERS[(Provider Config)]
WORKSPACES[(Workspace Config)]
end
STORE --- PROVIDERS
STORE --- WORKSPACES
classDef primary fill:#2374ab,stroke:#2374ab,color:#fff
classDef secondary fill:#ff7e67,stroke:#ff7e67,color:#fff
classDef store fill:#95b8d1,stroke:#95b8d1,color:#fff
class GW,GS,GR primary
class CLI,AI_APP secondary
class STORE,PROVIDERS,WORKSPACES store
The diagram shows the main components of the YAMCP system:
- CLI Commands: User interface for managing servers and workspaces
- McpGateway: Core component that coordinates the Gateway Server and Router
- GatewayServer: Handles communication with AI Apps via stdio transport
- GatewayRouter: Manages connections to configured MCP servers
- Logger: Provides consolidated logging for all components
- Store: Manages configuration for providers and workspaces
- MCP Servers: Both local (stdio) and remote (SSE) servers that provide MCP services
🪪 Security Audits
Comments
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.
