We have a multiple on prem network file locations.
What is the best way to setup Kinetic for Cloud operation to access these file directories via Application Studio apps and Kinetic Function C# Code blocks?
Thanks
Mike Ruoppoli
What types of files?
What kind of processing of files are you trying to do in Application studio and functions?
Depending on your situation you might have to get all your files in the cloud too.
If you are just parsing a file or two, they can be uploaded to Epicor to parse.
Links to local file attachments work if you have edge agent, I believe.
Hey Evan,
We are heavily customized with Classic Customizations.
We have multiple Network Mapped Drives that contain Project folders for Customer Information, Order Specific Information (Specs / POs etc), Template Job folders with PDF Drawings and DXF files.
We produce heavily engineered products, meaning we are not a widget shop.
Application studio functions would open and read different types of files, Spreadsheets, DXF files, PDFs that we combine into Packages currently using iTextSharp which i realize will go away when we go to the cloud.
I understand that we have to start using Sandbox.IO instead of System.IO for our functions.
That being said unfortunately we don’t just have a few files to upload.
When you say “Depending on your situation you might have to get all your files in the cloud too.” What does that look like?
How does SharePoint interface with Kinetic? Is this a viable option for our situation?
Hmm. You could go sharepoint or ECM, I don’t use them personally but I know there are lots of people on these forums using them so hopefully someone will chime in for their applicability.
Or you might want to keep the PDF merging/creation as a separate tool that lives on your local network that is triggered or polls epicor to look for merging tasks.
Would be curious to hear @klincecum or @Mark_Wonsil thoughts on this one.
Epicor has an integration natively for attachments. This can be leveraged as is, or worked with directly with some limitations to read / write files directly with code.
Epicor also includes the dlls for direct c# communication with sharepoint, and I have used that to write some more advanced features for reading and writing more broadly. I’m willing to share it and it’s not that complicated.
The dlls for MS Graph are also available.
Even direct http api access to the old or new apis for sharepoint is not that difficult, especially in the age of AI.
On prem storage with a SaaS Epicor is certainly possible, but not ideal. You can live with the limitations of attachments or uploading, or you can write some middleware with like Azure Relay Hybrid Connections, or a Minimal Api or the like.
The sky is the limit if you’re willing to work for it, or pay for it lol. 
My recommendation is unless you have just some wild requirements, make a plan to move these files to Sharepoint, and map out a plan for how you want to or Need to work with it. Then you can write out a detailed plan of the workflows you want, and we can advise you on how to make that happen.
We have 100s of thousands of files. We would never go back to local storage. We have similar engineering documents and such as you describe. It’s all doable, but for real advice, get a plan together, and ask specific questions.
I guess it depends on the company’s vision. If all you want is to keep doing what you’re doing, then I like Kevin’s suggestion to install the Kinetic Edge Agent on every machine that needs access to local file shares. For obvious security reasons, the browser operates in a sandbox to limit access to local resources. The Edge Agent is a local web service, which browsers can access, and it performs file access on behalf of the browser.
If you’re IT vision includes utilizing AI on all these files, then you might have to make some choices. You can run a local AI desktop product that has access to the local file shares and the online model. You can try to run a local LLM that indexes your local shares.
If you don’t want to use the Edge Agent, you can move the files to SharePoint or blob storage where you can grant access to LLMs. You can also use ECM and Prism - I think that was announced. 
The quickest solution is the Edge Agent. It’s the next step that will require some future planning to know the next best path.