Hello FME AI members,I have been using FME Flow MCP Server and MCP tools with Goose Desktop for PoC on Generative AI capabilities using Geospatial Data.The FME Workspaces I have authored now have a single SHP file being read with few attributes exposed as User Parameters (used in the FME Flow MCP Tool as well) and a MCP Writer. Can I have multiple FME Features read in a single workspace? Can I perform any GIS functions in the FME Workflow (using FME transformers) for example to check Point feature within an Area feature and use this workspace as a FME MCP tool for use by GPT to validate a query like for example a new utility service customer/ subscriber wants to know if his/her house falls can be serviced by an utility service company. The prospective customer provides his/her house address (a parameter in my FME Workspace), which is geocoded using FME to create a point, which is then overlaid using PointOnAreaOverlayer transformer in a FME Workspace. If the _overlaps count is not zero, then the response by the GPT should be serviceable. Can this be achieved? If yes, can you guide me?ThanksSRG

Question

FME Flow MCP and Geospatial Capabilities

+28

Hello FME AI members,

I have been using FME Flow MCP Server and MCP tools with Goose Desktop for PoC on Generative AI capabilities using Geospatial Data.

The FME Workspaces I have authored now have a single SHP file being read with few attributes exposed as User Parameters (used in the FME Flow MCP Tool as well) and a MCP Writer. Can I have multiple FME Features read in a single workspace?
Can I perform any GIS functions in the FME Workflow (using FME transformers) for example to check Point feature within an Area feature and use this workspace as a FME MCP tool for use by GPT to validate a query like for example a new utility service customer/ subscriber wants to know if his/her house falls can be serviced by an utility service company. The prospective customer provides his/her house address (a parameter in my FME Workspace), which is geocoded using FME to create a point, which is then overlaid using PointOnAreaOverlayer transformer in a FME Workspace. If the _overlaps count is not zero, then the response by the GPT should be serviceable. Can this be achieved? If yes, can you guide me?

Thanks

SRG