This topic describes the features of and billing information for Video Moderation Version 2.0.
I. Introduction to Video Moderation Version 2.0
Feature introduction
The Video Moderation Version 2.0 service identifies content or elements in ApsaraVideo VOD or live streaming videos that violate online content regulations, disrupt platform order, or negatively affect the user experience. This service lets you reuse previously configured Image Moderation Version 2.0 and Audio Moderation Version 2.0 services. With the Video Moderation Version 2.0 service from Content Moderation, you can perform further moderation or administration actions on video frame and audio content based on the rich risk labels and confidence scores returned by the API. You can customize these actions based on industry-specific standards or your platform's content administration rules.
Version comparison
Compared to Video Moderation 1.0, Video Moderation Version 2.0 provides more threat types, richer risk labels, and more flexible configuration features in the console.
Comparison item | Video Moderation Version 2.0 | Video Moderation 1.0 |
Default ingest endpoints | 50 | 20 |
Default QPS | 100 calls/second | 50 calls/second |
Default supported video size | 500 MB | 200 MB |
Supported threat detection scope | Video frames:
Note Video Moderation Version 2.0 integrates the Image Moderation Version 2.0 service. For more information about the Image Moderation Version 2.0 service, see Service description. Video audio:
Note Video Moderation Version 2.0 integrates the Voice Moderation Version 2.0 service. For more information about the Voice Moderation Version 2.0 service, see Service description. | Video frames:
|
Console features |
| Supports check item settings |
Billing | Billing is based on video frames and video audio (optional). Fee = Number of video snapshots × Number of services + Video duration × Unit price for video audio Frames are billed separately for each of the following business scenarios (optional, with pricing consistent with Image Moderation Version 2.0): General-purpose baseline check Video audio is billed based on video length, with a 10% discount compared to Voice Moderation Version 2.0. | Billing is based on frame risk scenarios (scene) and video audio (optional). Fee = Number of video snapshots × Number of risk scenarios + Video duration × Unit price for voice moderation Frames are billed separately for each of the following risk scenarios (optional, with pricing at 1.8 times that of Image Moderation 1.0):
Video audio is billed based on video length, with pricing consistent with Voice Moderation 1.0. |
Service description
The following table describes the services available in Video Moderation Version 2.0.
Service (service) | Detection Content | Scenarios |
Video File Detection (videoDetection_global) | Detects whether a video file contains violations in its frames or audio. | Detects non-compliant or unsuitable content in a video file. You can perform this check on all video files intended for public network access. |
Video File Detection (Large Model Edition) (videoDetectionByVL_global) | Uses the large model image moderation service to detect whether a video file contains violations in its frames or audio. Note This service is available in the Singapore region and supports 10 ingest endpoints by default. | Detects non-compliant or unsuitable content in a video file. You can configure large model image moderation rules. The service supports 10 ingest endpoints by default. You must control the number of calls. |
Live Video Stream Moderation (liveStreamDetection_global) | Detects whether a live video stream contains violations in its frames or audio. | Detects non-compliant or unsuitable content in a live video stream. You can perform this check on all live video streams intended for public network access. |
Live Video Stream Moderation (Large Model Edition) (liveStreamDetectionByVL_global) | Uses the large model image moderation service to detect whether a live video stream contains violations in its frames or audio. Note This service is available in the Singapore region and supports 10 ingest endpoints by default. | Detects non-compliant or unsuitable content in a live video stream. You can configure large model image moderation rules. The service supports 10 ingest endpoints by default. You must control the number of calls. |
II. Billing
Pay-as-you-go
When you activate the Video Moderation Version 2.0 service, the default billing method is pay-as-you-go. You are not charged if you do not call the service. The billing for the Video Moderation Version 2.0 API is detailed below.
Moderation Type | Supported Business Scenarios | Unit Price |
Video Frame Detection (General-purpose Edition) (image_standard,video_image_standard) |
| USD 0.6/1,000 calls Note You are charged once for each call to any of the business scenarios on the left. Billing is based on actual usage. For example, 100 calls to General-purpose baseline check cost USD 0.06. |
Video Frame Detection (Premium Edition) (image_advanced,video_image_advanced) |
| USD 1.2/1,000 calls Note You are charged once for each call to any of the business scenarios on the left. Billing is based on actual usage. For example, 100 calls to the large and small model fusion image moderation service cost USD 0.12. |
Video Audio Moderation (General-purpose Edition) (video_standard) |
| USD 8.1/1,000 minutes, which is equivalent to USD 0.486/hour. |
The pay-as-you-go usage of Content Moderation Version 2.0 is metered and billed once every 24 hours. In the billing details, the moderationType field corresponds to the moderation types listed above. For more information, see Bill Details.
III. Usage instructions
Integration instructions
You can integrate Video Moderation Version 2.0 in the following two ways:
Use a software development kit (SDK) to integrate the service. For more information, see Video Moderation Version 2.0 SDK and Integration Guide.
Use an API to integrate the service. For more information, see Video File Moderation Version 2.0 API.
Console operation guide
When you use the service for the first time, you must modify the video moderation configuration in the console.
You can modify video moderation policies, configure different policies for different business needs, view call results, and query usage.