About EPICS-MCP-Server
EPICS-MCP-Server is a Python MCP server for working with EPICS process variables, intended for environments that monitor or control hardware and software parameters. It provides tools to read and set values and retrieve PV details over standard input and output.
Key features
- 01Read process variable values
- 02Set process variable values
- 03Retrieve detailed PV information
- 04Communicate over standard input and output
Use cases
- 01Check monitored device parameters
- 02Adjust hardware or software parameters
MCP support
The MCP server exposes tools for reading and setting EPICS process variables and retrieving their details, communicating over stdio.
Configuration
Local server config is the same for every client. Copy it into your client’s MCP config file.
We haven’t collected a package or remote endpoint for this server. Follow the instructions in the project README.
Tools
Local servers expose tools only once running on your machine, and the maintainer has not provided a tool list.
Details
From the project READMEEPICS-MCP-Server
Overview
-
The EPICS MCP Server is a Python-based server designed to interact with EPICS (Experimental Physics and Industrial Control System) process variables (PVs). It provides a set of tools to retrieve PV values, set PV values, and fetch detailed information about PVs. The server is built using the mcp framework and communicates over stdio, making it suitable for integration into larger control systems or workflows.
-
This tool is particularly useful in environments where EPICS PVs are used for monitoring and controlling hardware or software parameters.
Features
- The EPICS MCP Server provides the following tools:
-
get_pv_value
- Create or update a single file in a repository
- Inputs:
pv_name(string): The name of the PV variable.
- Returns: A JSON object containing the status (
successorerror) and the retrieved value or an error message.
-
set_pv_value
- Set a new value for a specified PV.
- Inputs:
pv_name(string): The name of the PV variable.pv_value(string): The new value to be set for the PV.
- Returns: A JSON object containing the status (
successorerror) and a confirmation message or an error message.
-
get_pv_info
- Fetches detailed information about a specified PV.
- Inputs:
pv_name(string): The name of the PV variable.
- Returns: A JSON object containing the status (
successorerror) and the detailed information about the PV or an error message.
Usage with Langchain
- To use this with Langchain, you must install the dependencies required for the project.
pip install -r requirements.txt
-
Langchain
server_params = StdioServerParameters(
command="python",
# Make sure to update to the full absolute path to your math_server.py file
args=["/path/server.py"],
)
-
EPICS
- Before using the EPCIS mcp server, you must successfully install EPCIS on your local machine, ensure that IOC can start normally, and verify that functions such as
caget,caput, andcainfoare working properly. For detailed installation instructions, please refer to https://epics-controls.org/resources-and-support/base/.
jiangyan@DESKTOP-84CO9VB:~$ softIoc -d ~/EPICS/DB/test.db
Starting iocInit
############################################################################
## EPICS R7.0.8
## Rev. 2025-02-13T14:29+0800
## Rev. Date build date/time:
############################################################################
iocRun: All initialization complete
epics>
jiangyan@DESKTOP-84CO9VB:~$ caget temperature:water
temperature:water 88
jiangyan@DESKTOP-84CO9VB:~$ caput temperature:water 100
Old : temperature:water 88
New : temperature:water 100
jiangyan@DESKTOP-84CO9VB:~$ cainfo temperature:water
temperature:water
State: connected
Host: 127.0.0.1:5056
Access: read, write
Native data type: DBF_DOUBLE
Request type: DBR_DOUBLE
Element count: 1
Test Result
- Mcp client:
async def run():
async with stdio_client(server_params) as (read, write):
async with ClientSession(read, write) as session:
# Initialize the connection
await session.initialize()
# Get tools
tools = await load_mcp_tools(session)
# Create and run the agent
agent = create_react_agent(model, tools)
agent_response = await agent.ainvoke({"messages": "To query the value of a PV (Process Variable) named temperature:water"})
return agent_response
)
- Result:
================================[1m Human Message [0m=================================
To query the value of a PV (Process Variable) named temperature:water
==================================[1m Ai Message [0m==================================
Tool Calls:
get_pv_value (call_vvbXwi51CyYUxEM0hcyvCFCY)
Call ID: call_vvbXwi51CyYUxEM0hcyvCFCY
Args:
pv_name: temperature:water
=================================[1m Tool Message [0m=================================
Name: get_pv_value
{
"status": "success",
"value": 88.0
}
==================================[1m Ai Message [0m==================================
The current value of the PV named `temperature:water` is 88.0.
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.