Interface VideoAnalysisRich

Rich video analysis result with full scene descriptions, summary, transcript, and optional RAG chunk references.

This is a richer version of the base VideoAnalysis type that includes SceneDescription scenes (with cut types, confidence, key frames), a generated summary, and optional RAG indexing metadata.

Example

const analysis: VideoAnalysisRich = await videoAnalyzer.analyze(request);

console.log(`${analysis.sceneCount} scenes in ${analysis.durationSec}s`);
for (const scene of analysis.scenes) {
  console.log(`[${scene.startSec}s-${scene.endSec}s] ${scene.description}`);
}
interface VideoAnalysisRich {
    durationSec: number;
    sceneCount: number;
    scenes: SceneDescription[];
    summary: string;
    fullTranscript?: string;
    ragChunkIds?: string[];
    metadata: Record<string, unknown>;
}

Properties

durationSec sceneCount scenes summary fullTranscript? ragChunkIds? metadata

Properties

durationSec

durationSec: number

Total video duration in seconds.

sceneCount

sceneCount: number

Number of scenes detected.

scenes

scenes: SceneDescription[]

Ordered list of all detected scenes with rich descriptions.

summary

summary: string

Overall summary of the video content, generated by an LLM from the scene descriptions and transcript.

Optional fullTranscript

fullTranscript?: string

Full transcript of all audio in the video, when transcription was enabled. Concatenation of all scene transcripts with timestamp markers.

Optional ragChunkIds

ragChunkIds?: string[]

IDs of RAG vector store chunks created from this analysis. Only populated when VideoAnalyzeRequestRich.indexForRAG was enabled.

metadata

metadata: Record<string, unknown>

Additional metadata about the analyzed video. Provider-specific information that doesn't fit into the structured fields above.

Settings

Member Visibility

Theme

On This Page

durationSecsceneCountscenessummaryfullTranscriptragChunkIdsmetadata