A folder on this machine
Connect a local folder or a mounted network drive to a meinGPT data source.
This is the most common case and the reason the Outpost exists: a folder sits on the machine (or is mounted there as a drive), and meinGPT should be able to search it without the files leaving the building.
Only with your own Outpost
Only an Outpost can read a folder on a machine. Without your own Outpost this source type does not exist — you connect your storage in meinGPT directly instead (SharePoint, Google Drive, S3 …).
It is two steps, in two places
The connection is made of two halves, and the order is not arbitrary:
| Where | What | |
|---|---|---|
| 1 | In meinGPT | Create a data source of type Other — name, permissions, release level. It gets a data source ID. |
| 2 | In the Outpost | Link that ID to a folder path on this machine. |
The data source in meinGPT does not know where the folder is — that exists only on the Outpost machine. The Outpost does not know who may search — that exists only in meinGPT.
Create the data source in meinGPT
Data sources → new data source → Other. Give it a name people will recognise in chat later, and set the permissions.
Then copy the data source ID — a roughly 25-character string like cm4u4ovte000108lbfzv695kn. The Outpost spells out the path to it right under the field: meinGPT → Data sources → your data source.
Connect it in the Outpost
Data tab → Add data source → Other.
Two fields:
| Field | What goes in |
|---|---|
| Data source ID | The ID from step 1. |
| Folder path | An absolute path on this machine — C:\Data\Quotes or /srv/documents. Choose folder opens the operating system's file dialog. |
Save creates the source; it appears as a card in the Sources list.
Let it index
While the Outpost is running, the sync starts by itself. Sync on the card triggers it immediately.
If the status bar at the bottom reads Stopped, nothing happens — press Start first. The Index tab shows the progress.
Network drives
A network drive works as a local folder if it is permanently mounted from the Outpost process's point of view:
- Windows: a UNC path (
\\server\share\folder) or a mapped drive. Careful: mapped drive letters are per user session. If the Outpost runs as a different account it will not seeZ:\— use the UNC path or the SMB source type instead. - Linux: a mount point from
/etc/fstabthat survives a reboot.
The folder must be readable by the account the Outpost runs as. It does not need write access — do not give it any.
What gets read, and what does not
The Outpost reads the folder including subfolders and picks up new, changed and deleted files on later runs.
What it cannot process, it skips — and that is the thing that decides answer quality. The Index tab shows which extensions are affected under Skipped by type:
Typical large bars and what they mean:
.zip— archives are not unpacked. If the quote is inside a ZIP, nobody finds it..msg— individually exported Outlook mails. For mailboxes there is IMAP.no extension,.exe,.dll,.pyc— no text inside, correctly skipped.- Images and video — no text that can be extracted as-is.
A folder you release is open to everyone who can see the data source
File and directory permissions do not travel. Anyone allowed to see this data source in meinGPT finds every document in it — including someone who could not open the folder in Explorer. Check that before you release it, not after.
When nothing gets indexed
- Status bar: does it read
Running? If not, press Start. - Data tab: does the card show the path you expect? A typo is not reported here — the folder is simply not found.
- Logs tab, filter Ingest: an unreadable folder looks exactly like an empty one in the logs, but the reason is printed there.
- Index tab: if every file ended up under Skipped, it is not an access problem but a format problem.