This project demonstrates the integration of LabView automated tools into the self-driving lab software IvoryOS. The ability to control IvoryOS through Claude with an MCP interface allows to control the LabView software through natural language and easily integrate intelligent experimental design.
After running the 'main.py' file, the TCP server used to control the LabVIEW software can be started by clicking Devices → Chamber → Start Server. Then, the LabVIEW software needs to be started and the connection can be ensured by clicking Is Connected in the prior Ivory OS menu.
Afterwards, the LabVIEW software can be controlled either directly from the web interface or alternatively through natural language using the Claude interface:
Claude AI ◄──(MCP)──► IvoryOS Platform ◄──(TCP)──► LabVIEW Control
-
Claude AI: Natural language processing and intelligent control
-
MCP: Middleware Control Protocol for LLM-to-Ivory communication
-
IvoryOS: Automation platform with workflow management
-
TCP: Communication protocol layer connecting IvoryOS to LabVIEW
-
LabVIEW Control: Interface to the PVD system hardware
Developed during the LLM Hackathon for Applications in Materials Science & Chemistry as part of Ivory Zhang’s team in the Hein Lab at The University of British Columbia, this repository builds on the MCP server and provides an extended implementation.
This project is licensed under the MIT License.
Note: LabVIEW itself and any NI toolkits remain subject to their own licensing terms.