HTTP

Overview

The HTTP Event Action allows ANSVIS to send AI task event data to an external system via an HTTP POST request. This action is configured per AI task and is triggered automatically when the AI task generates an event (e.g., detection, counting, recognition, etc.).

Typical integrations:

Send event result to CMS / MES
Notify central monitoring server
Push AI metadata to analytics platform
Integrate with third-party automation

Target Server

URL: The endpoint that will receive the event data.

https://yourserver.com/api/event/receive

Method: All event data is sent in the POST request body.

POST

Media Options (optional)

ANSVIS can attach event-related media to the HTTP request.

Append Video

To use this feature, Cloud storage must be configured in system settings.

A short video clip is generated from camera frames around the event time.
The video is uploaded to the configured cloud storage.
The HTTP request includes a URL pointing to the uploaded video.
The Video Configurations dialog allows users to configure video display setting

Append Image

Image Size selection:

Cloud Image Storage: Image is uploaded to cloud storage and the POST request includes the image URL.
Base64 format: Image is converted to Base64 format with selected resolution and Base64 image is included directly in the POST body.

Add Object Overlays to Images If enabled, the bounding boxes and detection labels are drawn on the image before sending. User can disable it if they want to handle the overlay from the external system using the event data.

Payload Format (Style)

The Style option defines the structure of the HTTP POST body.

Available styles include:

Standard
Standard-Verbose
Custom Event Data
Serialized Custom Format
PLC Server
Fully Custom
Fully Serialized Custom

Standard Style

Uses the predefined ANSVIS event data structure.

You can refer to the Example Post Data on the right-side panel for standard style data (JSON structure). Each rule type and pipeline option will have different example.

Custom Style

Allows users to define a custom JSON structure.

Users can use the Configure Custom Format dialog to design the request body, including:

Insert dynamic variables to automatically replaced with actual values (e.g., camera ID, event name, confidence score) when the event occurs.
Rename fields
Control data hierarchy
Match external system format requirements

ANSVIS supports dynamic variables using the following syntax: !VariableName!. These variables are replaced with actual values when the event is triggered.

Variable Formatting Guidelines

Wrap variables with ! ! when using them in payload
Use quotation marks " " if the expected value is a string (Some variables are automatically formatted as string by ANSVIS)
Do not use quotation marks for numeric values
Refer to the Example Output Preview to verify final JSON formatting before deployment.

Core Event Variables

Variable	Type	Description	Availability
`EventID`	Integer	Unique identifier of the event	Always
`EventName`	String	Name of the triggered event	Always
`EventTime`	String (ISO 8601)	Event timestamp in readable format	Always
`EventUnixTime`	Integer	Event timestamp in Unix format	Always
`DGID`	Integer	Task group ID	System dependent
`i`	Integer	Internal event index or sequence counter	System dependent

Camera Variables

Variable	Type	Description	Availability
`CameraID`	Integer	Unique camera identifier	Always
`CameraName`	String	Camera display name	Always

Media Variables

These variables are available when image or video attachment is enabled.

Variable	Type	Description	Availability
`IMAGE`	Array[String]	Standard event image structure	When Base64 mode enabled
`AWSURL[]`	Array[String]	Uploaded image URLs	When cloud upload enabled
`AWSVIDEOURL`	String	Uploaded event video URL	When cloud upload enabled

AI Result Variables

Variable	Type	Description	Availability
`OverallConfidentScore`	Float	Overall confidence score of the event	AI Task
`DetectedObjectWithAttributes`	Object	Full detection result with associated attributes	AI Task
`DATA`	Object	Standard detection result structure	AI Task
`DDATA`	Object	Detailed detection data structure	AI Task
`ALPRDATA`	Object	Standard License Plate Recognition result structure	ALPR Task
`Rule[0].ROIs`	Object	Coordinate of Region of Interest if exists
`Rule[0].DetectionData[0].Detections[0].BoundingBox`	Float	Coordinate of detected object's bounding box
`Rule[0].DetectionData[].Name`	String	Class name of detected object
`Attribute[<Key>]`	String	Extracted attribute text	When attribute exists
`Attributes[<Key>]`	Array[String]	List of extracted attribute key	When multiple attributes exist

The Rule[...] variable are the one that uses a JSON query of the data.
- [0] specifies a specific item in an array.
- Empty brackets [] indicate all elements in the array.
<Key> must be replaced with the actual attribute name from JSON data, example:
- Text: License plate number
- color: License Plate Color
- Gender: Face gender

JSON data format:

{
  "EventName":"Vehicle Intrusion",
  "Detections":[
    {
      "RuleType":"Stable Object Appearing",
      "DetectionData":[
        {
          "Name":"motorbike",
          "ModelClassID":{"ModelID":785800007,"Class":2},
          "ModelClassName":{
            "Model":"B-IN_ANS_VehicleDetection(GPU) v1.0",
            "Class":"motorbike"
          },
          "DeploymentGroupID":1992040219,
          "DeploymentGroupName":"Vehicle Intrusion",
          "CameraID":1992040208,
          "CameraName":"Road",
          "CameraAddress":"BS_263_Daytime_1.mp4",
          "TrackID":1,
          "Detections":[
            {
              "Time":"2026-03-03T15:26:25.431Z",
              "BoundingBox":{
                "Left":1450,
                "Top":450,
                "Right":1655,
                "Bottom":797
              },
              "Score":0.871037,
              "Image":10,
              "ModelClassID":{"ModelID":785800007,"Class":2},
              "ModelClassName":{
                "Model":"B-IN_ANS_VehicleDetection(GPU) v1.0",
                "Class":"motorbike"
              },
              "ExtraInformation":"",
              "Attributes":[
                {
                  "Key":"Location",
                  "Value":"[1561,643,1610,680]"
                },
                {"Key":"Score","Value":"0.94"},
                {"Key":"Text","Value":"67B277363"},
                {"Key":"color","Value":"white"}
              ]
            }
          ],
          "Result":"5",
          "ExtraVisionResult":"",
          "TriggeredGroup":"",
          "UserInfo":{
            "UserID":"",
            "UserName":"",
            "GroupIDs":[],
            "GroupNames":[]
          },
          "Attributes":[
            [
              "ALPR",
              [
                ["Location","[0.09,0.11,0.57,0.32]"],
                ["Score","0.94"],
                ["Text","67B277363"],
                ["color","white"]
              ]
            ]
          ]
        }
      ],
      "PipelineOption":"ALPR pipeline",
      "ROIs":[
        {
          "ROIMatch":"Centre Point",
          "ROIPoints":[
            {"x":199,"y":515},
            {"x":573,"y":84},
            {"x":1743,"y":262},
            {"x":1698,"y":962},
            {"x":199,"y":990}
          ],
          "Option":"Inside ROI",
          "Name":"Rule 1",
          "OriginalImageSize":2546,
          "CustomModelROI":false,
          "RoiType":"Generic",
          "LineGap":0
        }
      ]
    }
  ],
  "Images":[
    {
      "ID":10,
      "Image":"20260303_220000/10.jpg",
      "ScaledFactor":1
    }
  ],
  "Video":"Video_20260303_222625.mp4"
}

Authentication (Optional)

If the external API requires authentication, enable Login Required.

ANSVIS will:

Send a login request to the configured Login URL.
Use the provided Login POST Data.
Extract the token from the login response.
Attach the token to the HTTP event request.

Login URL: The authentication endpoint of the external system.

https://yourserver.com/api/login

Login POST Data: JSON data required by the authentication API.

{
  "username": "your_user",
  "password": "your_password"
}

Token Type Name: Defines how the token is attached in the Authorization header.

Bearer //Example

Token Name: The field name in the login response containing the token.

accessToken //Example

Skip Trigger When Cloud Image Upload Failed

This option controls whether the HTTP event is sent using an Asynchronous (Async) or Synchronous (Sync) execution model when cloud media upload is enabled.

Disabled (Default)

Use this mode when real-time event delivery is more important than guaranteed media availability.

Execution Model: Asynchronous Mode (Async)
HTTP event is sent immediately after the AI event occurs.
Media upload runs in parallel.
Even if media upload fails, the HTTP request is still sent.

Enabled

Use this mode when the external system requires a valid and accessible media link for every event.

Execution Model: Synchronous Mode (Sync)
ANSVIS waits for media upload to complete.
HTTP event is sent only after successful upload.
If upload fails, the event is not sent.

Important Notes

Synchronous (Sync) mode may increase event latency.
This setting applies only when using cloud media upload (image/video).
Does not affect Base64 image mode.

Send Test Trigger

After finishing HTTP configuration, you can use the Send Test Trigger button to sends sample event data to the configured URL

This will helps verify: Connectivity, Authentication, Payload compatibility

Debug Mode

If enabled:

Logs HTTP request details.
Logs server response.
Useful for integration troubleshooting.

caution

Should be disabled in production environments.

Overview​

Target Server​

Media Options (optional)​

Append Video​

Append Image​

Payload Format (Style)​

Standard Style​

Custom Style​

Authentication (Optional)​

Skip Trigger When Cloud Image Upload Failed​

Disabled (Default)​

Enabled​

Send Test Trigger​

Debug Mode​