The Videos screen lists the videos your agent uses as references when it answers, and is where you add new ones from a file, YouTube or an online video link. Found in the agent sidebar under Knowledge β†’ Videos. Clicking a video opens its detail page.

Videos List

New VideoOpens the Add New Video flow, starting with the choice of video type.
Search VideosShows or hides the search panel below the buttons.
Sort BySorts the list by Code, Title, State or Type, in Ascending or Descending order.
Grid View / List ViewSwitches between thumbnail cards and a compact list. Your choice is remembered.
Video cardShows the thumbnail, a state badge, the title, the type and code (for example "YouTube - V01"), the author and how many chunks are active. Click it to open the detail page.
Load more dataLoads the next page of videos.

Search Videos

TitleFinds videos by title.
CodeFinds videos by code.
AuthorLimits the list to one registered author.
TypeLimits the list to User Uploaded, YouTube or Online Video.
StateLimits the list to one state (for example Active or Failed).
ApplyRuns the search.
ResetClears every search field.

Choose Video Type

Upload VideoUpload a video file from your computer.
YouTube VideoUse a YouTube link.
Online VideoUse a direct link to a video file on the web.

Add New Video: Source and Processing

The selected type is shown at the top of the form. The details cards appear after you pick a file (Upload Video), after Get Metadata succeeds (YouTube Video), or right away (Online Video).

FileUpload Video only. The video file to upload (required). It is stored temporarily and deleted after training finishes.
Video GuidelinesUpload Video only. Expands the allowed file types and other tips.
Video SourceYouTube Video and Online Video. The video URL (required). For YouTube, click Get Metadata to fill in the details and open the rest of the form.
Ingestion ModeHow the transcript is split into chunks (required). Options: Semantic or Skimming. See the Ingestion Modes on the Documents page.
... buttonOpens Ingestion Mode Parameters for the selected mode (content size and overlap for Semantic; training model, summarization and section sizes for Skimming).
LanguagesUp to 4 languages used for the transcription. Default: English (US) and Indonesian.

Video Details

TitleThe main name of the video (required).
AuthorA registered author, shown as code and name (required). Authors are managed under Citations β†’ Authors.
CodeYour own unique identifier for the video (required).
Thumbnail SourceChoose URL (paste an HTTPS image link and click Preview) or Upload (pick an image file).
PermissionsPermission codes that restrict this video to authorized users only.
Video SourceUpload Video only, and on the detail page. A link to the source or video location.
Save / CancelSave creates the video and starts processing.

Video Detail

Video StateThe current state of the video and the number of active chunks.
View ChunksOpens Chunks filtered to this video.
Chunking ProcessThe chunking status, with a View Log button that opens the Chunking Log.
Video PreviewPlays the video.
EditingThe details can only be edited while the video is Inactive.

Detail Actions

The buttons in the page header depend on the video's state.

DeactivateShown when the video is Active. Stops the agent from using it.
ActivateShown when the video is Inactive. Puts it back into use.
RetrainShown when Inactive. Lets you choose a new ingestion mode and parameters, then reprocesses the video.
Retrain as NewAvailable in retrain mode. Processes a copy of the video as a new video.
DeleteShown when Inactive. Deletes the video after confirmation.
Cancel ChunkingShown while the video is processing. Stops the processing.
Retry ChunkingShown when processing was Cancelled or Failed. You confirm by typing the video title.
Purge VideoShown for Deleted, Failed or Cancelled videos. Removes it permanently after you type the video title.

Related guide: Videos