Request API

Submit a video stream for real-time content moderation to detect regulatory risks and business-specific content in both captured frames and audio segments.

Frame detection identifies: political content, pornography, advertising, violence & terrorism, and other regulatory risks. It can also recognize faces, logos, flora & fauna, and other business-specific content based on your use case.

Audio detection identifies: political content, pornography, advertising, and other regulatory risks. It can also recognize gender, voice timbre, minors, and other business-specific content based on your use case.

API Description

Submit video stream information for moderation. Once stable stream pulling begins, detection results are continuously sent to the specified callback URL.

Requirements

Item	Specification
Protocol	HTTP or HTTPS
Method	POST
Encoding	UTF-8
Format	All request and response parameters use JSON

Supported Protocols

Standard stream URLs currently support RTMP, RTMPS, HLS, HTTP, and HTTPS protocols, including FLV and M3U8 formats.

Callback Mechanism

When the user receives a push result and returns an HTTP status code of 200, the push is considered successful. Otherwise, the system retries (up to the maximum retry count). Retry intervals in seconds: [5, 10, 15, 20, 25, 30, 35, 40, 45, 50, 55, 60]. After 12 failed attempts, no further retries are made.

Stream Pull Retry Mechanism

To prevent stream pull failures caused by network issues, the ISHUMEI video stream service has a built-in retry mechanism:

Standard streams / ZEGO / TRTC / Volcano streams: 12 retries total, each lasting 5 minutes, with intervals of [5, 10, 15, 20, ..., 60] seconds. For example, ISHUMEI first attempts to pull the stream for 5 minutes continuously. If unsuccessful, it waits 5 seconds then pulls for another 5 minutes. If still unsuccessful, it waits 10 seconds then pulls for another 5 minutes, and so on.
Agora streams: No retry. The connection is closed after a 5-minute pull timeout.

Timeout

Recommended timeout: 7s
Internal processing timeout is 3s with one automatic retry. Normal response time is within 100ms.

Request

Request URL

Cluster	Request URL	Supported Products
Shanghai Video Stream	`http://api-videostream-sh.fengkongcloud.com/videostream/v4`	Chinese Video Stream
Singapore Video Stream	`http://api-videostream-xjp.fengkongcloud.com/videostream/v4`	Chinese Video Stream, English Video Stream, Arabic Video Stream

Request Parameters

Parameter	Type	Required	Max Length	Description
`accessKey`	string	Yes	20	Company key for authentication, provided by ISHUMEI when the service is activated.
`eventId`	string	Yes	64	Event identifier. The value must be agreed upon with ISHUMEI in advance.
`appId`	string	Yes	64	Application identifier. This field is strictly validated and the value must be agreed upon with ISHUMEI in advance.
`imgType`	string	No	64	Regulatory detection types for video stream frames. At least one of `imgType` or `imgBusinessType` is required. See Image Detection Types.
`audioType`	string	No	64	Regulatory detection types for video stream audio. At least one of `audioType` or `audioBusinessType` is required. See Audio Detection Types.
`imgBusinessType`	string	No	128	Business detection types for video stream frames. At least one of `imgType` or `imgBusinessType` is required. See business label types for available values.
`audioBusinessType`	string	No	128	Business detection types for video stream audio. At least one of `audioType` or `audioBusinessType` is required. See Audio Business Types.
`imgCallback`	string	Yes	1024	Image callback URL. Frame capture detection results from the video stream are sent to this URL.
`audioCallback`	string	No	1024	Audio callback URL. Audio segment detection results from the video stream are sent to this URL. Required when audio detection is needed.
`data`	object	Yes	—	Request data content. Maximum size: 1 MB. See `data` Object.

Image Detection Types

Combine multiple types with underscores (e.g., POLITY_QRCODE_ADVERT).

Value	Description
`POLITY`	Political content detection
`EROTIC`	Pornographic & sexually suggestive content detection
`VIOLENT`	Violence, terrorism & prohibited content detection
`QRCODE`	QR code detection
`ADVERT`	Advertising detection
`IMGTEXTRISK`	Image text violation detection

Audio Detection Types

Combine multiple types with underscores (e.g., POLITY_EROTIC).

Value	Description
`POLITY`	Political content detection
`EROTIC`	Pornographic content detection
`ADVERT`	Advertising detection
`BAN`	Prohibited content detection
`VIOLENT`	Violence & terrorism detection
`DIRTY`	Abusive language detection
`ADLAW`	Advertising law violation detection
`MOAN`	Moaning detection
`AUDIOPOLITICAL`	Top leader voiceprint detection
`ANTHEN`	National anthem detection
`BANEDAUDIO`	Prohibited songs detection
`NONE`	Do not detect audio

Audio Business Types

Combine multiple types with underscores. To detect timbre, singing, or language, you must also include GENDER.

Value	Description
`SING`	Singing detection
`LANGUAGE`	Language detection (Chinese, English, Cantonese, Tibetan, Uyghur, Korean, Mongolian, Other)
`MINOR`	Minor detection
`GENDER`	Gender detection
`TIMBRE`	Voice timbre detection
`VOICE`	Voice attributes
`AUDIOSCENE`	Audio scene detection
`AGE`	Age detection
`APPNAME`	App name detection

`data` Object

Parameter	Type	Required	Max Length	Description
`streamType`	string	Yes	—	Video stream type. See Stream Types.
`tokenId`	string	Yes	64	User account identifier. Pass the user ID for risk detection of spam, advertising, and other behavioral dimensions.
`url`	string	No	600	URL of the standard video stream to be moderated.
`anchorName`	string	No	—	Anchor/host name. Typically used for manual review purposes.
`audioDetectStep`	int32	No	—	Audio moderation step size. Integer from 1–36. A value of 1 skips one 10-second audio segment, 2 skips two, and so on. When not set, all audio content is moderated.
`detectFrequency`	int32	No	—	Frame capture interval in seconds (1–60). Decimals are rounded down; values less than 1 are treated as 1. Default: 3 seconds.
`detectStep`	int32	No	—	Frame image detection step size. Only one frame per step is actually detected. Value ≥ 1. When not set, all captured frames are moderated.
`deviceId`	string	No	128	ISHUMEI device fingerprint identifier, generated by the ISHUMEI SDK for user behavior analysis.
`gender`	string	No	—	User gender. Suggested values: `male`, `female`, `ambiguity`.
`imgBusinessDetectStep`	int32	No	—	Image business label detection step size. Only one frame per step is detected for `imgBusinessType`. Value ≥ 1. Default: 1 (all frames are checked for business labels).
`imgCompareBase`	string	No	1024	Base image URL for face comparison. Required when `businessType` includes `FACECOMPARE`. Supported formats: jpg, jpeg, png, webp, gif, tiff, tif, heif. Recommended minimum resolution: 256×256. Animated image formats are not supported for base images.
`ip`	string	No	64	Client public IP address for IP-based user behavior analysis.
`lang`	string	No	—	Language type for text detection in captured frames and audio segments. Default: `zh`. See Supported Languages.
`level`	int32	No	—	User level. Different interception strategies can be configured per level. See User Levels.
`liveCover`	string	No	—	Live stream cover image. Typically used for manual review purposes.
`liveTitle`	string	No	—	Live stream title. Typically used for manual review purposes.
`receiveTokenId`	string	No	64	Message receiver's `tokenId`. Alphanumeric string with underscores and hyphens, up to 64 characters.
`returnAllImg`	int32	No	—	Controls which frames are returned in callbacks. `0` (default): return only non-pass risk level frames. `1`: return all risk level frames.
`returnAllText`	int32	No	—	Controls which audio results are returned in callbacks. `0` (default): return only non-pass risk level audio segments and text. `1`: return all risk level audio segments and text.
`returnFinishInfo`	int32	No	—	Stream end callback notification. `0` (default): do not send end notification. `1`: send end notification when moderation finishes; callback includes `statCode`.
`returnPreAudio`	int32	No	—	Whether to return the previous audio segment. `1`: `preAudioUrl` contains a 20-second audio link (previous 10s + current 10s) when the current segment is rejected. `0`: do not return previous segment info.
`returnPreText`	int32	No	—	Whether to return the previous audio segment text. `1`: `content` contains 20 seconds of text (previous 10s + current 10s) when the current segment is rejected. `0`: do not return previous segment text.
`room`	string	No	64	Live room / game room ID. Different strategies can be applied per room.
`streamName`	string	No	64	Video stream name. Displayed in the dashboard. Recommended to provide.
`agoraParam`	object	No	—	Agora recording parameters. Required when `streamType` is `AGORA`. See `agoraParam`.
`aliParam`	object	No	—	Alibaba Cloud recording parameters. Required when `streamType` is `ALI`. See `aliParam`.
`trtcParam`	object	No	—	Tencent TRTC recording parameters. Required when `streamType` is `TRTC`. See `trtcParam`.
`volcParam`	object	No	—	Volcano Engine recording parameters. Required when `streamType` is `VOLC`. See `volcParam`.
`zegoParam`	object	No	—	ZEGO recording parameters. Required when `streamType` is `ZEGO`. See `zegoParam`.
`extra`	object	No	—	Extra parameters.
`extra.passThrough`	object	No	1024	Client pass-through field. ISHUMEI does not process this field; it is returned as-is with the result.
`acceptLang`	string	No	—	Language for returned labels. `zh` (default): Chinese. `en`: English.

Stream Types

Value	Description
`NORMAL`	Standard stream URL. Supports RTMP, RTMPS, HLS, HTTP, HTTPS protocols.
`AGORA`	Agora moderation.
`TRTC`	Tencent TRTC moderation.
`ZEGO`	ZEGO moderation.
`VOLC`	Volcano Engine moderation.
`ALI`	Alibaba Cloud moderation.

💰
When using an RTC SDK recording solution, additional recording fees may be incurred on the RTC provider side. Consult the relevant RTC provider for specific pricing.

Supported Languages

Value	Language
`zh`	Chinese (default)
`en`	English
`ar`	Arabic
`hi`	Hindi
`es`	Spanish
`fr`	French
`ru`	Russian
`pt`	Portuguese
`id`	Indonesian
`de`	German
`ja`	Japanese
`tr`	Turkish
`vi`	Vietnamese
`it`	Italian
`th`	Thai
`tl`	Filipino
`ko`	Korean
`ms`	Malay
`auto`	Automatic language detection (contact ISHUMEI to enable interception standards)

User Levels

Value	Description
`0`	Lowest-level user (e.g., newly registered, completely inactive, or level-0 users)
`1`	Lower-level user (e.g., low activity or low-level users)
`2`	Mid-level user (e.g., moderately active or mid-level users)
`3`	Higher-level user (e.g., highly active or high-level users)
`4`	Highest-level user (e.g., paying users, VIP users)

`agoraParam` Object

Required when streamType is AGORA.

Parameter	Type	Required	Max Length	Description
`appId`	string	Yes	64	Application identifier provided by Agora.
`channel`	string	Yes	64	Channel name provided by Agora.
`channelProfile`	int32	No	32	Channel mode. `0` (default): Communication (any user can speak freely). `1`: Live broadcast (host and audience roles).
`enableH265Support`	boolean	No	—	Whether to support H.265 video stream recording. `false` (default): do not support. `true`: support H.265.
`enableIntraRequest`	boolean	No	—	Whether to enable keyframe requests. Default: `true` (improves audio/video experience on weak networks). Set to `false` to enable seeking in single-stream recordings. When `true`, single-stream recordings cannot seek to specific positions.
`subscribeMode`	string	No	—	Subscription mode. `AUTO`: automatically subscribe to all streams in the room (default). `UNTRUSTED`: only subscribe to users in `untrustedUserIdList`. `TRUSTED`: subscribe to all users except those in `trustedUserIdList`.
`token`	string	No	64	Authentication token for high-security scenarios. See Agora documentation for generation details. Set the token validity period longer than the channel duration. The maximum Agora token validity is 24 hours. For channels lasting longer than 24 hours, set `returnFinishInfo` to `1` and re-submit the channel with a new token when receiving an end callback (`statCode` = 1) due to token expiration.
`uid`	int32	No	64	32-bit unsigned integer. Required when `token` is provided — must match the user ID used to generate the token. This UID must not belong to an actual user in the room.
`trustedUserIdList`	array	No	—	Trusted user list. Effective when `subscribeMode` is `TRUSTED`. Cannot be empty. ISHUMEI will not subscribe to streams from these users. Comma-separated UID array (e.g., `[1,2]`), up to 17 users.
`untrustedUserIdList`	array	No	—	Untrusted user list. Effective when `subscribeMode` is `UNTRUSTED`. Cannot be empty. ISHUMEI will only subscribe to streams from these users. Comma-separated UID array (e.g., `[1,2]`), up to 17 users.

`aliParam` Object

Required when streamType is ALI.

Parameter	Type	Required	Max Length	Description
`room`	string	Yes	64	Room ID. Must exactly match the `channelID` used to generate the token. The service pulls and records streams per room. `room` is the unique identifier — duplicate `room` values will not trigger duplicate stream pulls.
`token`	string	Yes	64	Token for joining the channel. See Alibaba Cloud documentation for generation details. A new token must be generated for each moderation request.
`userId`	int32	No	32	Alibaba Cloud user account identifier.

`trtcParam` Object

Required when streamType is TRTC.

Parameter	Type	Required	Max Length	Description
`sdkAppId`	int32	Yes	64	SDK App ID provided by Tencent.
`strRoomId`	string	Yes	128	Room number. Only allows letters (a-zA-Z), digits (0-9), underscores, and hyphens. When both `strRoomId` and `roomId` are provided, `roomId` takes priority.
`userId`	string	Yes	32	User ID assigned to the recording client. Only allows letters (a-zA-Z), digits (0-9), underscores, and hyphens. Max 32 characters.
`userSig`	string	Yes	128	Authentication signature for the recording `userId`, equivalent to a login password.
`appScene`	int32	Yes	1	Application scenario. `0`: video call (default). `1`: video live broadcast. See Tencent documentation.
`demoSences`	int32	Yes	—	Recording type. `2`: separate stream recording. `4`: mixed stream recording.
`roomId`	int32	No	10	Room number (1–4294967294). Either `roomId` or `strRoomId` is required. When both are provided, `roomId` takes priority. A maximum of 8 users can be moderated per room.

`volcParam` Object

Required when streamType is VOLC.

Parameter	Type	Required	Max Length	Description
`appId`	string	Yes	64	Application identifier provided by Volcano Engine.
`roomId`	string	Yes	128	Room number.
`token`	string	Yes	64	Authentication signature for the recording `userId`, equivalent to a login password.
`userId`	string	Yes	32	User ID assigned to the recording client. Only allows letters (a-zA-Z), digits (0-9), underscores, and hyphens. Max 32 characters.

`zegoParam` Object

Required when streamType is ZEGO.

Parameter	Type	Required	Max Length	Description
`roomId`	string	Yes	64	ZEGO room number.
`tokenId`	string	Yes	64	ZEGO authentication token (`identify_token`). See ZEGO documentation for generation details. `tokenId` is a unique identifier — a new token must be generated for each moderation request.

Response

Response Parameters

Parameter	Type	Required	Description
`requestId`	string	Yes	Unique ISHUMEI request identifier.
`code`	int32	Yes	Response code. See Response Codes.
`message`	string	Yes	Response message corresponding to the `code`. See Response Codes.
`detail`	object	No	Detail information.
`detail.dupRequestId`	string	No	Duplicate `requestId`. Returned when `errorcode` is `1001` (duplicate stream push). Use this `dupRequestId` to call the close stream API if the original `requestId` was not received.
`detail.errorcode`	int32	No	`1001`: duplicate stream push.

Response Codes

Code	Message	Description
`1100`	Success	The request completed successfully.
`1901`	QPS limit exceeded	The request rate limit has been exceeded.
`1902`	Invalid parameters	One or more request parameters are invalid.
`1903`	Service failure	An internal service error occurred.
`1904`	Stream count limit exceeded	The maximum number of concurrent streams has been reached.
`9101`	Unauthorized operation	The provided `accessKey` does not have permission for this operation.

Stream Segment Callback Parameters

ℹ️
Parameters other than code, message, and requestId are only guaranteed to be returned when code is 1100.

Parameter	Type	Required	Description
`requestId`	string	Yes	Unique ISHUMEI request identifier.
`code`	int32	Yes	Response code. See Response Codes.
`message`	string	Yes	Response message corresponding to the `code`. See Response Codes.
`statCode`	int32	No	Callback status code. `0`: moderation result callback. `1`: stream end result callback.
`contentType`	int32	No	Distinguishes between audio and image callbacks. `1`: image callback. `2`: audio callback.
`auxInfo`	object	No	Auxiliary information. Contains the `passThrough` field from the request `data.extra`.
`auxInfo.passThrough`	object	No	Client pass-through field returned as-is.
`frameDetail`	object	No	Frame image risk details. Returned when risky frames exist or `returnAllImg=1`. See `frameDetail` Object.
`audioDetail`	object	No	Audio segment risk details. Returned when risky segments exist or `returnAllText=1`. See `audioDetail` Object.
`tokenProfileLabels`	array	No	Account attribute labels. Returned only when `tokenId` is provided and the labeling service is enabled. See Token Labels.
`tokenRiskLabels`	array	No	Account risk labels. Returned only when `tokenId` is provided and the labeling service is enabled. See Token Labels.

`frameDetail` Object

Parameter	Type	Required	Description
`imgUrl`	string	Yes	URL of the captured frame image.
`riskLevel`	string	Yes	Risk level. `PASS`: normal, `REVIEW`: suspicious, `REJECT`: violation.
`riskLabel1`	string	Yes	Level 1 risk label. Returns `normal` when `riskLevel` is `PASS`.
`riskLabel2`	string	Yes	Level 2 risk label. Empty when `riskLevel` is `PASS`.
`riskLabel3`	string	Yes	Level 3 risk label. Empty when `riskLevel` is `PASS`.
`riskDescription`	string	Yes	Risk description. Returns "Normal" when `riskLevel` is `PASS`. Format: "Level 1: Level 2: Level 3". Returns "Hit custom list" when a user-defined list is matched. For reference only — do not use for programmatic logic.
`allLabels`	array	Yes	List of all risk labels. See Frame `allLabels`.
`auxInfo`	object	Yes	Frame auxiliary information. See Frame `auxInfo`.
`riskDetail`	object	Yes	Risk detail information. See Frame `riskDetail`.
`businessLabels`	array	No	Business label list. See Frame `businessLabels`.

Frame `allLabels`

Each element in the allLabels array:

Parameter	Type	Required	Description
`riskLevel`	string	No	Risk level: `PASS`, `REVIEW`, or `REJECT`.
`riskLabel1`	string	No	Level 1 risk label.
`riskLabel2`	string	No	Level 2 risk label.
`riskLabel3`	string	No	Level 3 risk label.
`riskDescription`	string	No	Risk description. For reference only — do not use for programmatic logic.
`probability`	float	No	Confidence score between 0 and 1. Higher values indicate greater confidence.
`riskDetail`	object	No	Risk detail information. See Frame `riskDetail`.

Frame `auxInfo`

Parameter	Type	Required	Description
`beginProcessTime`	int32	Yes	Processing start time (13-digit timestamp).
`finishProcessTime`	int32	Yes	Processing end time (13-digit timestamp).
`detectType`	int32	No	Whether the frame was actually detected. Only returned when `detectStep` is set. `1`: frame was detected. `2`: frame was not detected.
`imgTime`	string	No	Absolute timestamp when the frame violation occurred.
`room`	string	No	Room number.
`similarityDedup`	int32	No	Returned only when the similar frame deduplication feature changes the outer `riskLevel` from reject/review to pass. Value `1` indicates deduplication was applied.
`strUserId`	string	No	User identifier for distinguishing violating users within a room. Unrelated to the request `userId`. Returned for: ZEGO room-based moderation, TRTC separate stream moderation, VOLC moderation, ALI moderation.
`userId`	int32	No	Agora user account identifier. Only present in separate stream scenarios. This is the actual user ID in the room, unrelated to the request `uid`.

Frame `riskDetail`

Parameter	Type	Required	Description
`riskSource`	int32	Yes	Risk source: `1000` (no risk), `1001` (text risk), `1002` (visual image risk).
`face_num`	int32	No	Number of faces detected.
`person_num`	int32	No	Number of persons detected.
`faces`	array	No	Names and positions of politically sensitive individuals in the image. Up to 10 entries (highest probability selected if more than 10). See Face Object.
`objects`	array	No	Detected objects/logos with names and positions. See Object Info.
`ocrText`	object	No	OCR text content. Present when `imgType` includes `IMGTEXTRISK` or `ADVERT`. Contains `text` (string): recognized text in the image.
`matchedLists`	array	No	Matched custom list information. Returned only when a custom list is hit. See Matched Lists.
`riskSegments`	array	No	High-risk content segments. Present when political, terrorism, prohibited, competitive, or advertising law content is detected. See Risk Segments.
`persons`	array	No	Person names and positions. When "person - multiple persons" label is hit, the array contains multiple elements (up to 10, highest probability selected). See Person Object.

Face Object

Parameter	Type	Required	Description
`id`	string	No	Identifier. The same person at the same position has the same ID across different labels. If the same person appears N times, N IDs are assigned.
`name`	string	No	Person name.
`face_ratio`	float	No	Face-to-image ratio (0–1). Higher values indicate a larger face proportion.
`probability`	float	No	Confidence score (0–1).
`location`	array	No	Face position coordinates `[x1, y1, x2, y2]` representing the top-left and bottom-right corners. Example: `[207, 522, 340, 567]` where 207=top-left X, 522=top-left Y, 340=bottom-right X, 567=bottom-right Y.

Object Info

Parameter	Type	Required	Description
`id`	string	No	Object/logo identifier. The same object at the same position has the same ID across different labels.
`name`	string	No	Object name.
`probability`	float	No	Confidence score (0–1).
`qrContent`	string	No	QR code URL detected in the image.
`location`	array	No	Object position coordinates `[x1, y1, x2, y2]` representing the top-left and bottom-right corners.

Matched Lists

Parameter	Type	Required	Description
`name`	string	No	Name of the matched list.
`words`	array	No	Sensitive word information from the matched list.
`words[].word`	string	No	The matched sensitive word.
`words[].position`	array	No	Position of the sensitive word.

Risk Segments

Parameter	Type	Required	Description
`segment`	string	No	High-risk content segment.
`position`	array	No	Position of the high-risk content segment (0-indexed).

Person Object

Parameter	Type	Required	Description
`id`	string	No	Identifier. The same person has the same ID across different labels. If the same person appears N times, N IDs are assigned.
`person_ratio`	float	No	Person-to-image ratio (0–1). Higher values indicate a larger person proportion.
`probability`	float	No	Confidence score (0–1).
`location`	array	No	Person position coordinates.

Frame `businessLabels`

Each element in the businessLabels array:

Parameter	Type	Required	Description
`businessLabel1`	string	Yes	Level 1 business label.
`businessLabel2`	string	Yes	Level 2 business label.
`businessLabel3`	string	Yes	Level 3 business label.
`businessDescription`	string	Yes	Business label description. Format: "Level 1: Level 2: Level 3".
`probability`	float	Yes	Confidence score (0–1).
`confidenceLevel`	int32	No	Confidence level (0–2). Higher values indicate greater confidence.
`businessDetail`	object	No	Business label details. May contain `face_num`, `person_num`, `faces`, `objects`, and `persons` with the same structure as described in Frame `riskDetail`.

`audioDetail` Object

Parameter	Type	Required	Description
`audioUrl`	string	Yes	URL of the audio segment.
`riskLevel`	string	Yes	Risk level: `PASS`, `REVIEW`, or `REJECT`.
`riskLabel1`	string	Yes	Level 1 risk label. Returns `normal` when `riskLevel` is `PASS`.
`riskLabel2`	string	Yes	Level 2 risk label. Empty when `riskLevel` is `PASS`.
`riskLabel3`	string	Yes	Level 3 risk label. Empty when `riskLevel` is `PASS`.
`riskDescription`	string	Yes	Risk description. For reference only — do not use for programmatic logic.
`vadCode`	int32	Yes	Whether this is a silent segment. `0`: silent. `1`: non-silent.
`allLabels`	array	Yes	List of all risk labels. See Audio `allLabels`.
`audioText`	string	No	Recognized text content of the current audio segment.
`content`	string	No	Audio text content. When `returnPreText` is `1` and the current segment is rejected, contains 20 seconds of text (previous 10s + current 10s). Otherwise, contains only the current segment text.
`preAudioUrl`	string	No	Previous audio segment URL. When `returnPreAudio` is `1` and the current segment is rejected, contains a 20-second audio link (previous 10s + current 10s). Otherwise, not returned.
`auxInfo`	object	No	Audio auxiliary information. See Audio `auxInfo`.
`businessLabels`	array	No	Business label list. See Audio `businessLabels`.

Audio `allLabels`

Each element in the allLabels array:

Parameter	Type	Required	Description
`riskLevel`	string	Yes	Risk level: `PASS`, `REVIEW`, or `REJECT`.
`riskLabel1`	string	Yes	Level 1 risk label. Returns `normal` when `riskLevel` is `PASS`.
`riskLabel2`	string	Yes	Level 2 risk label. Empty when `riskLevel` is `PASS`.
`riskLabel3`	string	Yes	Level 3 risk label. Empty when `riskLevel` is `PASS`.
`riskDescription`	string	Yes	Risk description. For reference only — do not use for programmatic logic.
`riskDetail`	object	No	Risk detail information. See Audio `riskDetail`.

Audio `riskDetail`

Parameter	Type	Required	Description
`riskSource`	int32	Yes	Risk source: `1000` (no risk), `1001` (text risk), `1002` (visual image risk), `1003` (audio voice risk).
`audioText`	string	No	Recognized text content of this segment.
`matchedLists`	array	No	Matched custom list information. See Matched Lists.
`riskSegments`	array	No	High-risk content segments. See Risk Segments.

Audio `auxInfo`

Parameter	Type	Required	Description
`audioStartTime`	string	Yes	Violation content start time (absolute time).
`audioEndTime`	string	Yes	Violation content end time (absolute time).
`beginProcessTime`	int32	Yes	Processing start time (13-digit timestamp).
`finishProcessTime`	int32	Yes	Processing end time (13-digit timestamp).
`room`	string	No	Room number.
`strUserId`	string	No	User identifier for distinguishing violating users within a room. Returned for: ZEGO room-based moderation, TRTC separate stream moderation, VOLC moderation, ALI moderation.
`userId`	int32	No	Agora user account identifier. Only present in separate stream scenarios.
`passThrough`	object	No	Client pass-through field returned as-is.

Audio `businessLabels`

Each element in the businessLabels array:

Parameter	Type	Required	Description
`businessLabel1`	string	Yes	Level 1 business label.
`businessLabel2`	string	Yes	Level 2 business label.
`businessLabel3`	string	Yes	Level 3 business label.
`businessDescription`	string	Yes	Business label description. Format: "Level 1: Level 2: Level 3".
`probability`	float	Yes	Confidence score (0–1).
`confidenceLevel`	int32	No	Confidence level (0–2). Higher values indicate greater confidence.
`riskDetail`	object	No	Risk detail information. See Audio `riskDetail`.

Token Labels

Both tokenProfileLabels and tokenRiskLabels share the same structure:

Parameter	Type	Required	Description
`label1`	string	No	Level 1 label.
`label2`	string	No	Level 2 label.
`label3`	string	No	Level 3 label.
`description`	string	No	Label description.
`timestamp`	int32	No	Label timestamp. 13-digit Unix timestamp in milliseconds.

Stream End Callback Parameters

Returned only when returnFinishInfo is 1.

Parameter	Type	Required	Description
`requestId`	string	Yes	Unique ISHUMEI request identifier.
`code`	int32	Yes	Response code. See Response Codes.
`message`	string	Yes	Response message corresponding to the `code`. See Response Codes.
`riskLevel`	string	Yes	Overall stream disposition recommendation at the end of moderation. `PASS`: normal, `REVIEW`: suspicious, `REJECT`: violation.
`statCode`	int32	Yes	Callback status code. `0`: moderation result callback. `1`: stream end result callback.
`contentType`	int32	Yes	Distinguishes callback end type. `1`: image moderation end callback. `2`: audio moderation end callback.
`pullStreamSuccess`	bool	Yes	Whether stream pulling was successful. `true`: success. `false`: failure (no frames were captured successfully).
`auxInfo`	object	Yes	Auxiliary information.
`auxInfo.streamTime`	int32	Yes	Stream moderation duration in seconds. Returned on the final callback after the stream ends. Represents the total moderation time (may differ from the actual stream duration if interval-based moderation is used).
`requestParams`	object	No	Returns all fields from the request `data` parameter. Returned when `contentType` is `2`.
`detail`	object	No	Detail information. Returned when `contentType` is `1`.
`detail.requestParams`	object	Yes	Returns all fields from the request `data` parameter.

Examples

Request Example

{
  "accessKey": "YOUR_ACCESS_KEY",
  "appId": "defaulttest",
  "audioBusinessType": "SING_LANGUAGE",
  "audioCallback": "http://www.example.com/callbackxxx",
  "audioType": "POLITY_EROTIC_ADVERT_MOAN",
  "data": {
    "detectFrequency": 10,
    "detectStep": 1,
    "extra": {
      "passThrough": {
        "passThrough1": "111",
        "passThrough2": "222",
        "passThrough3": "333"
      }
    },
    "ip": "123.171.34.4",
    "lang": "zh",
    "returnAllImg": 1,
    "returnAllText": 1,
    "returnPreAudio": 1,
    "returnPreText": 1,
    "room": "5e1854a6a0a79d0001a09bc3",
    "streamType": "NORMAL",
    "tokenId": "123",
    "url": "http://rtmp.example.com/live/3637778raLSXdOdu.flv"
  },
  "eventId": "VIDEOSTREAM",
  "imgBusinessType": "BODY_FOOD_3CPRODUCTSLOGO",
  "imgCallback": "http://www.example.com/callbackxxx",
  "imgType": "POLITY_EROTIC_ADVERT"
}

Response Example

{
  "code": 1100,
  "message": "Success",
  "requestId": "66fb85e3149bb9e13d6c72161cc6c6cf"
}

Frame Callback Example

{
  "auxInfo": {
    "passThrough": {
      "passThrough1": "111",
      "passThrough2": "222",
      "passThrough3": "333"
    }
  },
  "code": 1100,
  "contentType": 1,
  "frameDetail": {
    "allLabels": [
      {
        "riskDescription": "Politics: Political: Political",
        "riskLabel1": "politics",
        "riskLabel2": "shezheng",
        "riskLabel3": "shezheng",
        "riskLevel": "REJECT"
      }
    ],
    "auxInfo": {
      "beginProcessTime": 1639825248361,
      "detectType": 1,
      "finishProcessTime": 1639825248809,
      "imgTime": "2021-12-18 19:00:48.375",
      "room": "5e1854a6a0a79d0001a09bc3"
    },
    "businessLabels": [],
    "imgUrl": "http://example.com/image/frame_1639825248361.jpg",
    "riskDescription": "Politics: Political: Political",
    "riskDetail": {
      "ocrText": {
        "text": "Sample OCR text content"
      },
      "riskSource": 1002
    },
    "riskLabel1": "politics",
    "riskLabel2": "shezheng",
    "riskLabel3": "shezheng",
    "riskLevel": "REJECT"
  },
  "message": "Success",
  "requestId": "1639825145166_vs130_1639825248361471656"
}

Audio Callback Example

{
  "audioDetail": {
    "allLabels": [
      {
        "riskDescription": "Politics: Top leader: Top leader",
        "riskLabel1": "politics",
        "riskLabel2": "yihaolingdao",
        "riskLabel3": "yihaolingdao",
        "riskLevel": "REJECT"
      }
    ],
    "audioText": "Recognized audio text content...",
    "audioUrl": "http://example.com/audio_segment.mp3",
    "auxInfo": {
      "audioEndTime": "2022-10-27 21:08:42",
      "audioStartTime": "2022-10-27 21:08:32",
      "beginProcessTime": 1666876123332,
      "finishProcessTime": 1666876123893,
      "room": "y1123413312ewe24sv2"
    },
    "businessLabels": [],
    "content": "Recognized audio text content including previous and current segments...",
    "preAudioUrl": "http://example.com/audio_segment_pre.mp3",
    "riskDescription": "Politics: Top leader: Top leader",
    "riskDetail": {
      "audioText": "Recognized audio text content...",
      "riskSource": 1001
    },
    "riskLabel1": "politics",
    "riskLabel2": "yihaolingdao",
    "riskLabel3": "yihaolingdao",
    "riskLevel": "REJECT"
  },
  "code": 1100,
  "contentType": 2,
  "message": "Success",
  "requestId": "y28f8a4f1264085b321f12223wqed1121retestpvvvvv44321we12_3",
  "statCode": 0
}

Stream End Callback Example

{
  "auxInfo": {
    "streamTime": 70
  },
  "code": 1100,
  "contentType": 1,
  "detail": {
    "requestParams": {
      "detectFrequency": 10,
      "detectStep": 1,
      "extra": {
        "passThrough": {
          "passThrough1": "111",
          "passThrough2": "222",
          "passThrough3": "333"
        }
      },
      "ip": "123.171.34.4",
      "lang": "zh",
      "returnAllImg": 1,
      "returnAllText": 1,
      "returnPreAudio": 1,
      "returnPreText": 1,
      "room": "5e1854a6a0a79d0001a09bc3",
      "streamType": "NORMAL",
      "tokenId": "123",
      "url": "http://rtmp.example.com/live/3637778raLSXdOdu.flv"
    }
  },
  "message": "Success",
  "pullStreamSuccess": true,
  "requestId": "5515ce1f9b474a6c4a3d79a8dfcaeaf4",
  "riskLevel": "PASS",
  "statCode": 1
}

API Description

Requirements

Supported Protocols

Callback Mechanism

Stream Pull Retry Mechanism

Timeout

Request

Request URL

Request Parameters

Image Detection Types

Audio Detection Types

Audio Business Types

data Object

Stream Types

Supported Languages

User Levels

agoraParam Object

aliParam Object

trtcParam Object

volcParam Object

zegoParam Object

Response

Response Parameters

Response Codes

Stream Segment Callback Parameters

frameDetail Object

Frame allLabels

Frame auxInfo

Frame riskDetail

Face Object

Object Info

Matched Lists

Risk Segments

Person Object

Frame businessLabels

audioDetail Object

Audio allLabels

Audio riskDetail

Audio auxInfo

Audio businessLabels

Token Labels

Stream End Callback Parameters

Examples

Request Example

Response Example

Frame Callback Example

Audio Callback Example

Stream End Callback Example

`data` Object

`agoraParam` Object

`aliParam` Object

`trtcParam` Object

`volcParam` Object

`zegoParam` Object

`frameDetail` Object

Frame `allLabels`

Frame `auxInfo`

Frame `riskDetail`

Frame `businessLabels`

`audioDetail` Object

Audio `allLabels`

Audio `riskDetail`

Audio `auxInfo`

Audio `businessLabels`