SurfSense vs NotebookLM: file chat and connectors

This Persian-language walkthrough uses SurfSense's hosted interface. The presenter contrasts its model choices and connectors with NotebookLM, and describes self-hosting on a computer or server; installation is not covered. The demonstration uses Google sign-in and cloud model choices, so it does not establish offline operation.

The presenter adds a YouTube camera setup guide and a PDF camera manual. The YouTube connector also accepts playlists. Google Drive is unavailable during the recording, so that connector is discussed rather than tested. The PDF upload enables Vision LLM to interpret images as well as text and uses basic processing. Processing takes several minutes in this example.

An initial summary request fails to identify a source because no document is selected. The presenter then checks the relevant source, summarizes the YouTube guide and adds the camera manual to the chat. A question about changing the camera's frame rate returns a menu path that the presenter says is correct. This selection step is the practical focus of the tutorial.

The closing section describes scheduled automation and agent tasks using connected sources. These receive an overview rather than an executed workflow. Model availability, processing options and connector status reflect the interface at the time of recording.