This topic describes the billable items for Intelligent Media Management.
Billable item pricing
Billable items for Intelligent Media Management (IMM) include: image intelligence, metadata management, media processing, document processing, and file processing.
Effective 11:00 on July 28, 2025 (UTC+8), Intelligent Media Management (IMM) will introduce charges for some currently free features and adjust pricing for some existing billable items. For more information, see Announcement on IMM billing adjustments.
Image intelligence
The following table details the pricing for image intelligence billable items.
Billable item | Description | Billing adjustment | Related API operations | Related x-oss-process operations | Price (CNY) | Unit |
ImageDetect | face detection | Consolidates charges formerly billed under |
|
| 0.20 | per 1,000 calls |
body detection | Moves from free public preview to general availability. | DetectImageBodies (image body detection) | image/bodies | |||
car detection | Moves from free public preview to general availability. | DetectImageCars | image/cars | |||
ImageLabel | image labeling | Consolidates charges formerly billed under | DetectImageLabels (image label detection) | image/labels | 1.00 | per 1,000 calls |
ImageFace | face image | Moves from free public preview to general availability. | CreateFacesSearchingTask (create a similar face image search task) | 0.20 | per 1,000 calls | |
ImageFaceClustering | face clustering |
| 50.00 | per 1,000 calls | ||
GenerateStory | story generation | CreateStory (create a story) | 50.00 | per 1,000 calls | ||
ImageMosaic | image mosaic | Moves from free public preview to general availability. | AddImageMosaic (add an image mosaic) | 0.05 | per 1,000 calls | |
ImageCropping | intelligent image cropping suggestion | DetectImageCropping (detect visually appealing crop areas) | image/crop,g_auto | 1.00 | per 1,000 calls | |
ImageQRCodes | image QR code detection | DetectImageCodes (image QR code detection) | image/codes | 0.80 | per 1,000 calls | |
ImageSplicing | image splicing | Moves from free public preview to general availability. | CreateImageSplicingTask (create an image splicing task) |
| per 1,000 calls | |
ImageToPDF | image to PDF conversion | Moves from free public preview to general availability. | CreateImageToPDFTask (create an image-to-PDF conversion task) | 50.00 | per 1,000 images | |
ImageScoring | image quality scoring | DetectImageScore (image quality scoring) | image/scoring | 0.30 | per 1,000 calls | |
LocationDateClustering | spatio-temporal clustering | CreateLocationDateClusteringTask (create a spatio-temporal clustering task) | Free of charge for a limited time | per 1,000 calls | ||
SimilarImageClustering | image clustering | CreateSimilarImageClusteringTask (create a similar image clustering task) | Free of charge for a limited time | per 1,000 calls | ||
Blindwatermark | blind watermarking for images |
|
| 0.70 | per 1,000 calls | |
ReverseGeocoding | reverse geocoding | DetectMediaMeta (get media file metadata) Note Charges apply when the media file contains geolocation information. | 1.00 | per 1,000 calls | ||
ImageTexts | image text recognition (OCR) | DetectImageTexts (image text recognition) | 50.00 | per 1,000 calls |
Metadata management
The following table details the pricing for metadata management billable items.
Billable item | Description | Related API operations | Related x-oss-process operations | Price (CNY) | Unit |
StandardQueryL0 | basic query |
| task/get | 0.005 | per 1,000 calls |
StandardQueryL1 | standard query |
| 0.10 | per 1,000 calls | |
StandardQueryL2 | advanced query |
| 0.50 | per 1,000 calls | |
MediaMeta | media information retrieval |
|
| 1.00 | per 1,000 calls |
SemanticAnalyze | semantic analysis | SemanticQuery (natural language query) | 3.50 | per 1,000 calls |
Media processing
The following table details the pricing for media processing billable items.
Billable item | Description | Related API operations | Related x-oss-process operations | Price (CNY) | Unit |
AudioCompress | audio transcoding | CreateMediaConvertTask (create a media transcoding task) |
| 0.00010 | per second of audio |
VideoCompressCopy | remuxing | CreateMediaConvertTask (create a media transcoding task) | 0.00010 | per second of video | |
VideoCompress264LD | CreateMediaConvertTask (create a media transcoding task) |
| 0.00036 | per second of video | |
VideoCompress264SD | CreateMediaConvertTask (create a media transcoding task) |
| 0.00050 | per second of video | |
VideoCompress264HD | CreateMediaConvertTask (create a media transcoding task) |
| 0.00090 | per second of video | |
VideoCompress2642K | CreateMediaConvertTask (create a media transcoding task) |
| 0.00200 | per second of video | |
VideoCompress2644K | CreateMediaConvertTask (create a media transcoding task) |
| 0.00450 | per second of video | |
VideoCompress265LD | CreateMediaConvertTask (create a media transcoding task) |
| 0.00150 | per second of video | |
VideoCompress265SD | CreateMediaConvertTask (create a media transcoding task) |
| 0.00250 | per second of video | |
VideoCompress265HD | CreateMediaConvertTask (create a media transcoding task) |
| 0.00500 | per second of video | |
VideoCompress2652K | CreateMediaConvertTask (create a media transcoding task) |
| 0.00800 | per second of video | |
VideoCompress2654K | CreateMediaConvertTask (create a media transcoding task) |
| 0.01600 | per second of video | |
MediaAnimation | video to animated image conversion | CreateMediaConvertTask (create a media transcoding task) | video/animation |
| per 1,000 frames |
ExtractSubtitleText | video text subtitle extraction | CreateMediaConvertTask (create a media transcoding task) | 1.50 | per 1,000 tracks | |
ExtractSubtitleImage | video image subtitle extraction | CreateMediaConvertTask (create a media transcoding task) | 0.10 | per minute | |
VideoFraming | video frame capture | CreateMediaConvertTask (create a media transcoding task) |
| 0.10 | per 1,000 frames |
VideoClassification | video label detection | CreateVideoLabelClassificationTask (create a video label classification task) | 50.00 | per 1,000 calls | |
LiveTranscoding | GenerateVideoPlaylist (generate a playlist for live transcoding) |
| 0.00010 | per call |
Document processing
The following table details the pricing for document processing billable items.
For projects created before December 1, 2023, online preview and online editing are billed per document open. For projects created on or after December 1, 2023, these services are billed per API call.
Billable item | Description | Related API operations | Related x-oss-process operations | Price (CNY) | Unit |
DocumentConvert | document conversion | CreateOfficeConversionTask (create a document conversion task) |
| 80.00 | per 1,000 calls |
document content extraction | ExtractDocumentText (extract document body text) | ||||
DocumentWebofficeEdit |
| doc/edit | 20.00 | per 1,000 calls | |
DocumentWebofficePreview |
| doc/preview | 10.00 | per 1,000 calls | |
DocumentWebofficeCachePreview | cached preview (Weboffice) |
| 7.00 | per 1,000 calls Important Billed per API call. |
File processing
The following table details the pricing for file processing billable items.
Billable item | Description | Related API operations | Related x-oss-process operations | Price (CNY) | Unit |
PointCloudCompress | point cloud compression | CreateCompressPointCloudTask (create a point cloud compression task) | pointcloud/compress | 0.20 | per 1,000 calls |
FileProcess | file compression | CreateFileCompressionTask (create a file compression task) | 0.0050 | per GB | |
archive decompression | CreateFileUncompressionTask (create an archive decompression task) | ||||
FilePreview | archive preview | CreateArchiveFileInspectionTask (create an archive preview task) | 0.050 | per TB |
APIs with multiple billable items
The
SemanticQueryoperation includes theStandardQueryL2andSemanticAnalyzebillable items.The
CreateFacesSearchingTaskoperation includes theImageDetectandImageFacebillable items.
Video transcoding
H.264 transcoding: The output video uses the H.264 codec.
H.265 transcoding: The output video uses the H.265 codec.
LD: The resolution of the output video is less than or equal to 640 x 480.
SD: The resolution of the output video is less than or equal to 1280 x 720.
HD: The resolution of the output video is less than or equal to 1920 x 1080.
2K: The resolution of the output video is less than or equal to 2560 x 1440.
4K: The resolution of the output video is less than or equal to 3840 x 2160.
Video transcoding is charged per second of the output video. For billing, the transcoding duration is rounded up to the next second.
Billing for document preview and editing
For projects created before December 1, 2023, billing for online document editing and document preview is based on the number of document opens, not API calls.
For projects created on or after December 1, 2023, billing is based on the number of API calls. To switch to the new billing model, create a new project.
In the API call-based billing model, each token from an API call is valid for a single user. If a token is reused, only the last user retains access, and the permissions for previous users are revoked.
Setting the
Permission.Readonlyparameter totruein theGenerateWebofficeTokenAPI call incurs charges for document preview. Setting it tofalseincurs charges for online document editing.The billing for a
RefreshWebofficeTokencall is determined by thePermission.Readonlysetting in the originalGenerateWebofficeTokencall. If this parameter was set totrue, the refresh is billed as document preview; otherwise, it is billed as online document editing.
Billing for LiveTranscoding
Billing components:
When you generate a playlist, you can set the
InitialTranscodeparameter to control the duration of initial transcoding. This action incurs LiveTranscoding fees.During playback, playing a non-transcoded TS file triggers a new transcoding task, which incurs LiveTranscoding fees.
Transcoding incurs fees for reading source video files from OSS and writing transcoded files to OSS. Reading video files from OSS for playback also incurs fees. For more information about OSS-related charges, see OSS billable items.
LiveTranscoding Count Unit (CU) calculation:
Video
The CU coefficient (eff) for the video codec is 0.3 for h264 and 1.8 for h265.
Calculation formula:
Ceiling (eff * Ceiling(Height/240) * Ceiling(Width/240) * Ceiling(FrameRate/30) + 1 ) * Ceiling(VideoStreamDuration)Audio
The CU coefficient (eff) is 0.3.
Calculation formula:
Ceiling(eff * Ceiling(AudioStreamDuration))
Billing rules: You are charged separately for each processed video stream and audio stream, which are specified by
TargetVideo.StreamandTargetAudio.Streamrespectively. The following examples demonstrate how LiveTranscoding fees are calculated.Example 1 (Generating a playlist without playback incurs no LiveTranscoding fees):
A user calls
GenerateVideoPlaylistfor a 38-minute video, generating an output with a resolution of 800x600, a frame rate of 30, and the h264 video codec. The initial transcoding duration is 0 seconds, andTranscodeAheaduses the default value. The video is not played.
Example 2 (Generating a playlist with pre-transcoding but without playback incurs LiveTranscoding fees only for the pre-transcoded portion):
A user calls
GenerateVideoPlaylistfor a 38-minute video, generating an output with a resolution of 800x600, a frame rate of 30, and the h264 video codec. The initial transcoding duration is 30 seconds, andTranscodeAheaduses the default value. The video is not played.Fees incurred:
LiveTranscoding (The number of CUs is calculated as follows):
Ceiling((0.3 * Ceiling(800/240) * Ceiling(600/240) * Ceiling(30/30) + 1 ) * (Ceiling(30)) + Ceiling(0.3 * Ceiling(30)) = 159 CUs
Example 3 (Generating a playlist and playing video segments incurs LiveTranscoding fees only for played segments):
A user calls
GenerateVideoPlaylistfor a 38-minute video, generating an output with a resolution of 800x600, a frame rate of 30, and the h264 video codec. The initial transcoding duration is 0 seconds, andTranscodeAheaduses the default value. The user plays the video using the generated M3U8 file, starts from the beginning and plays to the 5-minute mark (with a default of 2 minutes transcoded ahead), then seeks to the 15-minute mark and plays to the end.Fees incurred:
LiveTranscoding (The number of CUs is calculated as follows):
Ceiling((0.3 * Ceiling(800/240) * Ceiling(600/240) * Ceiling(30/30) + 1) * (Ceiling((5+2)*60) + Ceiling((38-15)*60)) + Ceiling(0.3 * Ceiling((5+2) * 60)) + Ceiling(0.3 * Ceiling((38-15) * 60)) = 9,540 CUs
Example 4 (If multiple users play the video, LiveTranscoding fees are charged only once for replayed segments):
A user calls
GenerateVideoPlaylistfor a 38-minute video, generating an output with a resolution of 800x600, a frame rate of 30, and the h264 video codec. The initial transcoding duration is 0 seconds, andTranscodeAheaduses the default value.User A plays the video from the beginning to the 5-minute mark and then stops playback.
User B starts playback at the 15-minute mark and plays to the end.
User C plays the video from the beginning to the end.
Fees incurred:
LiveTranscoding (The number of CUs is calculated as follows):
Ceiling((0.3 * Ceiling(800/240) * Ceiling(600/240) * Ceiling(30/30) + 1) * Ceiling(38*60)) + Ceiling(0.3 * Ceiling(38 * 60)) = 12,084 CUs
Terms
Width: Width of the output video.
Height: Height of the output video.
FrameRate: The video frame rate.
VideoStreamDuration: The duration of the video stream.
AudioStreamDuration: The duration of the audio stream.
eff: The CU coefficient.
Ceiling(x): Returns the smallest integer that is greater than or equal to x.
Operators and billable items
When you create a metadata index by binding an OSS bucket or by calling the IndexFileMeta or BatchIndexFileMeta operations, executing the operators specified in Workflow templates and operators incurs data processing fees, index storage fees, and OSS request fees. Object Storage Service (OSS) charges these request fees. For more information, see OSS Request fees. The following table describes the relationship between operators and billable items.
Operator | Billable item | Charged by |
OSSMeta operator | OSS | |
MIME operator | Free of charge | N/A |
FaceDetection operator | ImageFace* | IMM |
LabelClassification operator (images) | ImageClassification* | IMM |
LabelClassification operator (videos) | VideoClassification | IMM |
ImageScoring operator | ImageScoring* | IMM |
ReGEO operator | ReverseGeocoding | IMM |
MediaMeta operator | MediaMeta | IMM |
EXIF operator | OSS | |
ExtractDocumentText operator | DocumentConvert | IMM |
ExtractImageEmbeddings operator | Free of charge for a limited time | IMM |
To support various image formats, IMM uses the image processing (IMG) capabilities of Object Storage Service (OSS) to perform image operations, such as format conversion and image resizing. OSS charges fees for these operations. For more information about these fees, see Data processing fees.
External request fees
When you use IMM to access OSS, you incur OSS request fees. For more information, see API operation calling fees.
Billing differences
The new version of Intelligent Media Management is no longer billed based on the QPS configured for a project (instance).
The new version is billed on an hourly basis only; daily billing is no longer available.
In the new version, requests that return a 4XX response status code are billable, including those with a parameter error or permission error.