Feature Request: API Commands, Camera Tasks, and Alert Popups

[imortezamosavi]

Hello iSpyConnect Support Team,

I am developing a plugin for iSpyConnect, and I would like to request some new features for the API commands and alert system.

Currently, I need the following capabilities:

1-Add a tag or task when adding an IP camera
Add support for specifying a tag or task when using the command/addCamera API command.
2-Add an API command to retrieve camera tasks
Add a new API command, such as command/getTask, that allows retrieving the tasks associated with cameras.
3-Improve the alert notification system
Add a new version of the alert functionality that displays the relevant camera frame/image in a popup window when an alert is triggered.

These features would be very useful for integrating iSpyConnect with external AI and monitoring systems.

Could you please let me know if these features are currently available, planned, or could be added to a future version?

Thank you for your support.

[ispysoftware]

Hello sure - could you expand on some of your asks - like i'm not really sure what you mean by "Add a tag or task when adding an IP camera" or "Add a new version of the alert functionality that displays the relevant camera frame/image in a popup window when an alert is triggered."

• this is already handled by
  https://www.ispyconnect.com/docs/agent/alerts?q=alerts#alert-display

so im not really sure what you're after.. maybe some mockups of what you are looking for would help.

[imortezamosavi]

Hi, thanks for your response. Let me clarify the three requests with a concrete example of how I am using Agent.

I am developing an external AI monitoring system that integrates with Agent. The external AI service processes the camera streams and performs different types of detection, such as cow detection, face recognition, fire detection, intrusion detection, etc.

1. Assign a task/tag to a camera when using addCamera

I would like to associate a task or tag with a camera when creating it through:

/command/addCamera

For example:

Camera 12 → task = cow_detection
Camera 13 → task = face_recognition
Camera 14 → task = fire_detection

So I would like to be able to specify something like:

/command/addCamera?
name=Barn Camera 12&
substream=...&
mainstream=...&
task=cow_detection

The purpose is to store metadata that tells my external AI service what processing should be performed for each camera.

If Agent already has an existing mechanism for storing custom metadata, tags, notes, or similar information for a camera, and this information can be accessed through the API, I would be happy to use that instead. I may simply be missing the existing functionality.

---

2. Retrieve the task/tag associated with cameras through the API

The second requirement is the ability to retrieve this information.

For example, my external AI service would like to query Agent and get something like:

{
    "camera_id": 12,
    "name": "Barn Camera 12",
    "task": "cow_detection"
}

The exact API design is not important.

I would simply need an API command that allows my external service to retrieve the task/tag/custom metadata associated with a camera.

For example, conceptually:

/command/getCamera

or:

/command/getTask

If this functionality already exists through another API command, please let me know which command I should use.

---

3. Automatically open the Alert Display when an AI detection occurs

For the third request, I now understand that Agent already has the Alert Display functionality.

The part I am looking for is the automatic opening/displaying of the alert.

According to the documentation, the user currently clicks an alert thumbnail to open the Alert Display and see the alert image, detection boxes, video preview, AI description, etc.

In my use case, the alert is generated by an external AI service.

For example:

Camera 12
    ↓
External AI
    ↓
Cow detected
    ↓
Detection frame captured
    ↓
Alert sent to Agent

When this happens, I would like Agent to automatically open/show the relevant Alert Display, without requiring the user to first find the alert and click its thumbnail.

In other words, I would like this behavior:

AI detects something
        ↓
Agent receives/creates alert
        ↓
Agent automatically opens Alert Display
        ↓
User immediately sees the alert image

I am not asking for a new Alert Display UI. I would like to reuse the existing Alert Display functionality.

My question is:

Does Agent currently provide an API command, event, URL, or another mechanism that allows an external application to programmatically open/show the Alert Display for a specific alert or camera?

If an external AI application can already create an alert through the API, I mainly need a way to tell Agent:

"A new alert has occurred for Camera X — immediately display/open this alert to the user."

If this functionality does not currently exist, then this is the feature I would like to request.

Overall workflow

The intended architecture is:

IP Cameras
     ↓
Agent
     ↓
External AI Service
     ↓
AI Detection
     ↓
Alert/Event
     ↓
Agent automatically opens Alert Display
     ↓
User immediately sees the detection

The goal is to integrate Agent with an external AI system while still using Agent's existing alert and display capabilities.

Thanks again for taking the time to clarify this. If any of these capabilities already exist in Agent, I would be very happy to use the existing API rather than requesting new features.

[ispysoftware]

Thanks for the detailed write-up - here's where each of these landed.

1. Custom metadata on addCamera

Agent already has a free-text Notes field on every camera (and microphone) - it just wasn't reachable from the API. In the next update, addCamera (and addOnvif) accept an optional notes parameter:

/command/addCamera?name=Barn Camera 12&substream=...&mainstream=...&notes=cow_detection
URL-encode the value like the other parameters. It's stored in the camera's Notes field, so be aware it's also visible and editable in the camera's settings in the UI - fine for tagging, but don't treat it as private storage. I deliberately kept it generic (notes) rather than adding a task concept, since it's just metadata as far as Agent is concerned.

2. Reading it back

getObjects (and getObject) now include a notes field in the JSON for each device, so your service can call /command/getObjects and get the id, name, type and notes for every camera in one request - that's your camera_id → task mapping.

3. Auto-opening the Alert Display

This one I'm not going to add, and it's a deliberate call rather than an oversight. An API command that forces UI open on every connected client is a bad primitive: anything with API access could take over every screen, a burst of detections would leave the UI constantly popping modals (and interrupting anyone mid-configuration), and with multiple people connected there's no single "the user" to show it to. Once a command like that exists, it also can't be un-shipped.

There are two existing mechanisms that get you the outcome without those problems:

Alert slot in a layout. When editing a layout you can add an alert slot - a cell that automatically switches to show the live feed of whichever camera is currently alerting. Since your external detections come in through the alert pipeline (/command/alert?ot=2&oid=X), a cow detection on camera 12 will put camera 12 live in that slot the moment the alert fires, with no clicking at all. For a monitoring-station setup this is exactly the "user immediately sees it" behavior you described, and it's the user's choice to run that layout.
Broadcast messages. The next update also cleans up the broadcast API: /command/broadcast?text=Cow detected on Barn Camera 12 shows a notification on every connected client immediately. Note there's no throttling on this path - if your AI fires frequently, rate-limit on your side.
So the flow would be: your service detects → calls /command/alert for camera X (which drives recording, actions, push notifications and the alert list) → optionally calls /command/broadcast for an instant on-screen message → and a layout with an alert slot displays the camera live automatically. The user stays in control of their own screen, which is where I want to keep that decision.

Note the /command/alert route requires Agent to be armed and alerts enabled on the device, and it respects the configured alert interval - your service should expect and handle the error responses for those states.

[imortezamosavi]

Thanks for the detailed explanation. That clears everything up.
I checked the Layout documentation and I now understand how the Alert Slot works.

[imortezamosavi]

Thanks for the clarification. I reviewed the Alert Slot documentation and tested/considered how it works, but unfortunately it doesn't cover the behavior I have in mind.

What I need is slightly different from the current Alert Slot behavior.

I want the camera feed to be live continuously, even when there is no active alert.

The desired behavior is:

1. The Live Feed is always visible in the layout.
2. When my AI detects an object and I trigger /command/alert for that camera, a popup/alert overlay should appear on top of the existing live feed.
3. The live video must continue playing underneath the popup.
4. When the alert ends, the popup should disappear, but the live feed should remain visible.
5. The screen should never turn black just because there is no active alert.

Currently, the Alert Slot appears to behave differently: it is black when there is no alert, shows the camera feed when an alert is triggered, and then becomes black again after the alert ends.

So I don't want the Alert Slot to control whether the live feed is displayed. I need a persistent Live Camera Feed + temporary Alert Popup/Overlay triggered by /command/alert.

Is there an existing layout/camera component that supports this behavior, or could the alert mechanism be changed so that it triggers an overlay on an already-running live feed instead of switching the slot on/off?

This would allow the normal monitoring view to remain visible at all times while still providing an immediate visual indication whenever my external AI detects an event.

[ispysoftware]

you can do this with the broadcast command -
/command/broadcast?text=xxx
pass whatever you want into that - it displays a popup in every attached UI

also usually you'd only have the alert slot in the corner of the layout - so the camera can still be completely visible in a grid on the same view but the alert slot only becomes active when the device is alerting.

[imortezamosavi]

Thanks. I understand the Alert Slot approach.

One thing I noticed, though: I'm currently using Agent DVR 7.7.8.0, and when I check the local Swagger API documentation, I don't see /command/broadcast listed under the available Commands.

You mentioned:

/command/broadcast?text=xxx

Could you clarify whether this endpoint is already available in 7.7.8.0 but simply not documented in Swagger, or whether it is available only in a newer build?

If it is already available in 7.7.8.0, could you please provide a quick example of how I should call it?

Also, regarding the alert indicator triggered by /command/alert / result.alert, is there any way to customize its position or style? Ideally, I would like the notification to appear as a centered overlay on the live camera feed while the video continues playing underneath.

[ispysoftware]

it's in latest version
it might work in 7780 - you'd need to try it

"Ideally, I would like the notification to appear as a centered overlay on the live camera feed while the video continues playing underneath."
no it needs it's own space - it's terrible UX to pop alerts over live video

the plugin examples show how to draw whatever you want on the live video stream though - you could just add an overlay in the plugin.

[imortezamosavi]

Hi Again,

I’m testing the Agent DVR API and noticed an issue with the broadcast command.

My Swagger UI shows:

Agent DVR Swagger/API version: 7.9.1.0
Application version: *7.8.9.0
Swagger specification: *OAS 2.0

When I call the broadcast endpoint:

curl -X 'GET'
'http://localhost:8090/command/broadcast?text=test'
-H 'accept: application/json'

the API responds successfully:

{"status":"ok"}

However, nothing appears in the Agent DVR UI. I tested this while the UI was open and connected.

For comparison, the alert command works correctly and the UI displays the alert:

curl -X 'GET'
'http://localhost:8090/command/alert?oid=2&ot=2&tag=fsce&message=%D8%B7%D8%B1%D8%B7'
-H 'accept: application/json'

So the API connection itself appears to be working, and the UI is able to receive and display alerts. The problem seems to be specific to the broadcast command.

Could you please confirm:

1. Is /command/broadcast supposed to display a popup/message in the Web UI in this version?
2. Is it supported in application version 7.8.9.0*?
3. Does broadcast require a specific UI/client connection or configuration?
4. Is there another API/event mechanism you recommend for triggering a custom popup or notification in the UI?

I’m specifically looking for a way to trigger a notification from a C# plugin without covering the live camera video.

Thanks.

[ispysoftware]

Sounds like you'll need to update the application version

…

On Tue, 11 Aug 2026, 6:57 pm Morteza Mousavi, ***@***.***> wrote: *imortezamosavi* left a comment (ispysoftware/AgentDVR-Plugins#8) <#8 (comment)> Image: Image (view on web) <https://github.com/user-attachments/assets/28006ab2-7893-4725-856f-fcc41cc3e327> Hi Again, I’m testing the Agent DVR API and noticed an issue with the broadcast command. My Swagger UI shows: Agent DVR Swagger/API version: 7.9.1.0 Application version: *7.8.9.0 Swagger specification: *OAS 2.0 When I call the broadcast endpoint: curl -X 'GET' 'http://localhost:8090/command/broadcast?text=test' -H 'accept: application/json' the API responds successfully: {"status":"ok"} However, nothing appears in the Agent DVR UI. I tested this while the UI was open and connected. For comparison, the alert command works correctly and the UI displays the alert: curl -X 'GET' ' http://localhost:8090/command/alert?oid=2&ot=2&tag=fsce&message=%D8%B7%D8%B1%D8%B7' -H 'accept: application/json' So the API connection itself appears to be working, and the UI is able to receive and display alerts. The problem seems to be specific to the broadcast command. Could you please confirm: 1. Is /command/broadcast supposed to display a popup/message in the Web UI in this version? 2. Is it supported in application version 7.8.9.0*? 3. Does broadcast require a specific UI/client connection or configuration? 4. Is there another API/event mechanism you recommend for triggering a custom popup or notification in the UI? I’m specifically looking for a way to trigger a notification from a C# plugin without covering the live camera video. Thanks. — Reply to this email directly, view it on GitHub <#8?email_source=notifications&email_token=AAGDKXIMBO2PGMCSRQBGYWL5JL3YNA5CNFSNUABFM5UWIORPF5TWS5BNNB2WEL2JONZXKZKDN5WW2ZLOOQXTKMRVGIZDGMRXGE2KM4TFMFZW63VHMNXW23LFNZ2KKZLWMVXHJLDGN5XXIZLSL5RWY2LDNM#issuecomment-5252232714>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAGDKXPGVA2KIRL6Y756UND5JL3YNAVCNFSNUABFKJSXA33TNF2G64TZHM2DCNJXGMZDSNRWHNEXG43VMU5TKMBZGYYDMMZZHAZKC5QC> . You are receiving this because you commented.Message ID: ***@***.***>