VideoObject

The VideoObject type represents a distinct video asset, not the page hosting it. Use @type: "VideoObject" to declare the video as a first-class entity. This distinction is critical for AI parsers that must separate the video from its HTML container.

• name: The title of the video
• description: A textual summary of the content
• thumbnailUrl: A URL pointing to a representative image frame
• contentUrl: The direct URL to the video file itself

Always provide a unique @id IRI to enable cross-referencing within a larger knowledge graph.

The duration property specifies the total length of the video using ISO 8601 duration format. This structured temporal data allows search engines to display video length directly in results and enables AI models to understand content scope.

• duration: "PT1M33S" represents 1 minute and 33 seconds
• uploadDate: The ISO 8601 date the video was published
• expires: Optional date when the video becomes unavailable

Accurate duration markup correlates with higher click-through rates in video-rich results.

The hasPart property links a VideoObject to Clip entities that define specific segments. Each Clip uses SeekToAction to specify a timestamp, enabling AI-generated key moment navigation in search results.

• Clip: Defines a named segment with startOffset and endOffset
• SeekToAction: An action pointing to a specific temporal offset via startOffset
• PotentialAction: Used at the top level to declare interactive seek targets

This markup directly powers the "Key Moments" feature in Google Search, allowing users to jump to specific sections identified by AI.

Distinguish between the video file and its player. The contentUrl points to the actual media file (MP4, WebM), while embedUrl points to a player page. This separation lets AI agents choose the appropriate resource for different contexts.

• contentUrl: Direct binary file URL for download or streaming
• embedUrl: URL of a player page suitable for iframe embedding
• encodingFormat: MIME type such as "video/mp4"

Providing both URLs ensures compatibility with crawlers that index files and those that render embedded players.

Link a transcript property to associate a text representation of spoken content. This textual data is critical for AI models performing semantic analysis, as it provides a searchable corpus of the video's dialogue.

• transcript: URL to a plain text or WebVTT file
• caption: URL to a closed caption file for accessibility
• inLanguage: BCP 47 language code of the audio track

Transcripts enable LLMs to retrieve and cite specific spoken passages, dramatically improving the video's visibility in generative search overviews.

The interactionStatistic property uses InteractionCounter to declare quantitative engagement metrics. This structured data signals content authority and popularity to ranking algorithms.

• InteractionCounter: Ties a metric to a specific interactionType
• WatchAction: Declares the number of views
• LikeAction: Declares the number of likes or positive reactions

Accurate engagement data, when verifiable, contributes to the entity's perceived authority within knowledge graphs.