Request scene summary

Available in VPC

Request a VLM-based scene summary for an index that has been fully indexed. Once the scene summary is complete, a natural language text summary will be generated for each scene.

Note

Note the following when requesting a scene summary.

Scene summaries are executed based on the index. You can only request scene summarization for indexes where indexing has been completed.
The initial analysis must be run on the entire scene. Once the full analysis is complete, you can perform a partial reanalysis by specifying specific scenes.
For a given asset, scene summary can only be run on a single index.
Editing metadata (such as person labeling or script editing) is restricted while scene summarization is in progress.

Request

This section describes the request format. The method and URI are as follows:

Method	URI
POST	/api/v1/workspaces/{workspace_name}/projects/{project_id}/assets/{asset_id}/indexes/{index_id}/scene-summaries

Request headers

For information about the headers common to all Media Intelligence APIs, see Media Intelligence request headers.

Request path parameters

The following describes the parameters.

Field	Type	Required	Description
`workspace_name`	String	Required	Workspace name
`project_id`	String	Required	Project ID See Get project list.
`asset_id`	String	Required	Media asset ID See Get media asset list.
`index_id`	String	Required	Index ID See Get index list.

Request body

You can include the following data in the body of your request:

Field	Type	Required	Description
`sceneIds`	Array<Integer>	Optional	List of scene IDs to be reanalyzed Available only after the initial full analysis has been completed If not entered, all scenes in the index will be analyzed. If entered, only the specified scenes will be reanalyzed, and only the summaries for those scenes will be updated (summaries for other scenes will remain unchanged).

Request example

The request example is as follows:

curl --location --request POST 'https://mi.apigw.ntruss.com/api/v1/workspaces/my-workspace/projects/1234/assets/5678/indexes/1001/scene-summaries' \
--header 'x-ncp-apigw-timestamp: {Timestamp}' \
--header 'x-ncp-iam-access-key: {Access Key}' \
--header 'x-ncp-apigw-signature-v2: {API Gateway Signature}' \
--header 'Content-Type: application/json' \
--data '{
    "sceneIds": [1, 3, 5]
}'

Response

This section describes the response format.

Response body

The response body includes the following data:

Field	Type	Required	Description
`code`	String	-	API processing result code
`message`	String	-	API processing result message
`result`	Object	-	Scene summary request information
`result.sceneSummaryId`	Integer	-	Scene summary request ID
`result.indexId`	Integer	-	Analysis target index ID
`result.targetSceneIds`	Array<Integer>	-	List of scene IDs to be analyzed `null` if it is a full analysis
`result.createdTime`	String	-	Analysis request creation date and time ISO 8601 format
`result.createdUserName`	String	-	Username of the user who requested the analysis

Response status codes

For information about the HTTP status codes common to all Media Intelligence APIs, see Media Intelligence response status codes.

Response example

The response example is as follows:

{
    "code": "0",
    "message": "success",
    "result": {
        "sceneSummaryId": 101,
        "indexId": 1001,
        "targetSceneIds": null,
        "createdTime": "2026-04-17T10:30:00",
        "createdUserName": "username"
    }
}

Documentation Index