# IND360 integration with MQTT

In this video and guide, we explore how **IND360** can be seamlessly integrated via MQTT with:

* **Node-RED:** A flow-based, low-code programming tool for wiring together hardware, APIs, and online services, ideal for rapid IIoT prototyping. \
  🔗 <https://nodered.org/docs/getting-started/local>
* **Ignition by Inductive Automation** is an industrial application platform for building SCADA, HMI, and IIoT solutions with unlimited clients and a web-based deployment model. \
  🔗 <https://inductiveautomation.com/downloads/ignition>

***

### 📺 Video Tutorial

{% embed url="<https://youtu.be/7Ml6OcBUadA>" %}

***

### 🔑 What You Will Learn in This Video

In this hands-on demo, you will learn:

* How to connect **IND360 with Node-RED and Ignition SCADA** using **MQTT**
* How to read **weight & status data** in real time
* How to send **commands** such as:
  * Zero scale
  * Tare scale
  * Preset tare
  * Clear tare
* How to log the feedback in the database

***

### Prerequisites

| Requirement           | Details                                      | Status     |
| --------------------- | -------------------------------------------- | ---------- |
| Mettler Toledo IND360 | With network access via Ethernet             | ● Required |
| MQTT Broker           | Shiftr (for testing) or Ignition Distributor | ● Required |
| Node-RED              | Installed locally or on server               | Use Case 1 |
| Ignition Gateway      | Standard edition v8.3.4+                     | Use Case 2 |
| Cirrus Link Modules   | MQTT Distributor + MQTT Engine               | Use Case 2 |

***

### 1.  MQTT Broker Setup

Before connecting the IND360, you need a running MQTT broker. For initial testing, we use **Shiftr,** an MQTT broker with a visual connection map.&#x20;

{% hint style="success" %}
For production use, Ignition's built-in broker is used.
{% endhint %}

{% stepper %}
{% step %}

### Download and install the Shiftr desktop app

Visit `https://www.shiftr.io/desktop` and install the broker locally.
{% endstep %}

{% step %}

### MQTT Broker parameters

Note the MQTT port address
{% endstep %}

{% step %}

### Verify port 1883 is open

Open PowerShell and run the commands below.

```powershell
# Verify Port is Open and Listening
netstat -ano | findstr :1883

# Find Which App Owns the Port
tasklist -ano | findstr <pid>

# Open port 1883 in Windows Firewall
New-NetFirewallRule -DisplayName "MQTT Broker" -Direction Inbound -Protocol TCP -LocalPort 1883 -Action Allow

```

{% hint style="warning" %}
**Important — Firewall**

If the IND360 cannot connect via IP address, the Windows Firewall is blocking port 1883. Run the PowerShell command above to open it.
{% endhint %}
{% endstep %}
{% endstepper %}

### 2.  Enabling MQTT on IND360

In this step, we need to enable the MQTT service on the IND360 device.

{% stepper %}
{% step %}

### Open the IND360 web interface

Navigate to `https://192.168.0.8` (replace with your IND360 IP) and login using `002`.
{% endstep %}

{% step %}

### Navigate to MQTT settings

Go to **Communication → Service** and find the MQTT section.
{% endstep %}

{% step %}

### Configure MQTT parameters

Fill in all fields as shown in the table below.

| Parameter   | Value          | Notes                             |
| ----------- | -------------- | --------------------------------- |
| Client ID   | `IND360`       | Unique identifier for this device |
| Host        | `192.168.0.24` | IP of your broker machine         |
| Port        | `1883`         | Standard MQTT port                |
| Topic       |                | Root topic prefix                 |
| Username    |                | Must match broker user            |
| Supervision | `Enabled`      | Enables keep-alive monitoring     |

{% hint style="info" %}
**Topic Structure**

The IND360 automatically appends its serial number and sub-paths to the root topic. For example: `ind360/devices/C249346600/scales/_/measurements`
{% endhint %}
{% endstep %}

{% step %}

### Click SET, then CONNECT

Status should change to **Connected**.
{% endstep %}
{% endstepper %}

***

### 3.  Testing with Shiftr

Use Shiftr's visual broker map to confirm the IND360 is publishing data before integrating with Node-RED or Ignition.

**Open the Shiftr dashboard to visualize the data flowing from the IND360 device**

{% embed url="<https://files.gitbook.com/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FLd2M9UNfMTnw9DjDdZJz%2Fuploads%2FvQCsuBulKx3R5NjHO0a4%2Fmqtt%20broker.mp4?alt=media&token=9f4b62f7-7759-4068-80fc-9d42dbfa457e>" %}

***

### Use Case 1: Node-RED Integration

Use Node-RED as a lightweight MQTT client to read live weight data and send commands to the IND360 via the Shiftr broker.

#### **Reading Weight Data**

Subscribe to the measurements topic to receive live weight data as a JSON array.

* ind360/devices/C249346600
* ind360/devices/C249346600/alarms
* ind360/devices/C249346600/scales/\_
* ind360/devices/C249346600/scales/\_/measurements

<figure><img src="/files/lamc447UeZ93zIdGpOG5" alt=""><figcaption></figcaption></figure>

```javascript
//function 21
msg.payload = msg.payload[0].gross + ' ' + msg.payload[0].uomCode;
return msg;

//function 22
msg.payload = msg.payload[0].net;
return msg;

//function 23
msg.payload = msg.payload[0].tare + ' ' + msg.payload[0].uomCode;
return msg;

```

#### Sending Commands <a href="#nodered-write" id="nodered-write"></a>

Each command follows the same pattern — an Inject node triggers a Function node that builds the payload, which is then published via an MQTT Out node.

<table><thead><tr><th width="116.3333740234375">Command</th><th width="153.333251953125">Description</th><th>Topic</th></tr></thead><tbody><tr><td>tare</td><td>Sets net weight to zero using current gross as tare offset</td><td><p><strong>Invoke:</strong></p><p><code>ind360/devices/C249346600/commands/tare/invoke</code></p><p><strong>Result:</strong><br><code>ind360/devices/C249346600/commands/tare/result</code></p></td></tr><tr><td>tare-clear</td><td>Removes the tare offset, gross equals net again</td><td><p><strong>Invoke:</strong></p><p><code>ind360/devices/C249346600/commands/tare-clear/invoke</code></p><p><strong>Result:</strong><br><code>ind360/devices/C249346600/commands/tare-clear/result</code></p></td></tr><tr><td>zero</td><td>Sets current gross weight as the new zero reference point</td><td><p><strong>Invoke:</strong></p><p><code>ind360/devices/C249346600/commands/zero/invoke</code></p><p><strong>Result:</strong><br><code>ind360/devices/C249346600/commands/zero/result</code></p></td></tr><tr><td>tare-preset</td><td>Applies a manually entered known tare value</td><td><p><strong>Invoke:</strong></p><p><code>ind360/devices/C249346600/commands/tare-preset/invoke</code></p><p><strong>Result:</strong><br><code>ind360/devices/C249346600/commands/tare-preset/result</code></p></td></tr></tbody></table>

<figure><img src="/files/TvC2xhAQQdBuYjKc4v9c" alt=""><figcaption></figcaption></figure>

**Tare Command — Function Node**

```javascript
msg.payload = {
    "$correlation": "TareTest1",   // Unique ID to track this request
    "$ttl": 42,                   // Time-to-live in seconds
    "responseTopic": null,         // Set null if no response topic needed
    "command": "tare",             // Command to execute
    "content": null               // Not required for tare
}
return msg;
```

**Clear Command — Function Node**

```javascript
msg.payload =
{
    "$correlation": "TareTest1",
    "$ttl": 42,
    "responseTopic": null,
    "command": "tare-clear",
    "content" : null
}
return msg;
```

**Zero Command — Function Node**

```javascript
msg.payload =
{
    "$correlation": "Zero-Test1",
    "$ttl": 42,
    "responseTopic": null,
    "command": "zero",
    "content" : null
}
return msg;
```

**Preset Tare Command — Function Node**

```javascript
msg.payload = {
    "$correlation": "Tare-PresetTest1",
    "$ttl": 42,
    "responseTopic": null,
    "command": "tare-preset",
    "content": {
        "value": "2.0",             // Preset tare value
        "unit": "Kg"               // Unit — must match scale config
    }
}
return msg;
```

***

### Dashboard

You can use the dashboard 2.0 to visualize your value as shown below. Feel free to use the NodeRed flow for a ready-to-use dashboard.&#x20;

<div align="left"><figure><img src="/files/MKBAbyVhPR2wwsV94z1L" alt="" width="314"><figcaption></figcaption></figure></div>

{% file src="/files/17hlvKsxADLURl6EfcRg" %}

***

### Use Case 2: Ignition Integration

Ignition acts as both the MQTT broker (via MQTT Distributor) and the subscriber (via MQTT Engine). No external broker needed.&#x20;

{% stepper %}
{% step %}

### Install Ignition

Sign up for an account, download, and install Ignition. <https://inductiveautomation.com/>
{% endstep %}

{% step %}

### Install MQTT Modules

Visit `https://inductiveautomation.com/downloads/third-party-modules/8.3.4` — Match the version to your Ignition installation.

**Install via Gateway → Platform → Modules**

<table><thead><tr><th width="175.33331298828125">Module</th><th width="437.0001220703125">Role</th><th>Port</th></tr></thead><tbody><tr><td>MQTT Distributor</td><td>Acts as the MQTT broker inside Ignition</td><td>1883</td></tr><tr><td>MQTT Engine</td><td>Subscribes and converts MQTT data to Ignition tags</td><td>—</td></tr><tr><td></td><td></td><td></td></tr></tbody></table>
{% endstep %}

{% step %}

### Configure MQTT Distributor

**Gateway → Connections → MQTT Distributor → Settings**

Enable the distributor and verify that the port is set to `1883`.

**Create broker users**

Go to Users and create two accounts:

| Username      | Password   | Used by                                              |
| ------------- | ---------- | ---------------------------------------------------- |
| `ind360`      | `ind360`   | IND360 device                                        |
| `ignition`    | `ignition` | MQTT Engine internal connection                      |
| `nodered`     | `nodered`  | (Optional) For Node-RED to subscribe and publishdata |
| {% endstep %} |            |                                                      |

{% step %}

### Configure MQTT Engine

**Gateway → Connections → MQTT Engine → Servers → Add Server**

| Field         | Value                     |
| ------------- | ------------------------- |
| URL           | `tcp://192.168.0.24:1883` |
| Server Set    | `Default Set`             |
| Username      | `ignition`                |
| Password      | `ignition`                |
| {% endstep %} |                           |

{% step %}

### Create Custom Namespace

{% hint style="info" %}
**Why a Custom Namespace?**

MQTT Engine's default namespaces (Sparkplug B and Elecsys) expect specific structured payload formats. The IND360 publishes plain JSON — so we must create a custom namespace that tells the Engine how to parse and tag the incoming data.
{% endhint %}

**Gateway → MQTT Engine → Namespaces → Custom → Create New**

| Field                | Value           | Notes                            |
| -------------------- | --------------- | -------------------------------- |
| Namespace Name       | `IND360`        | Unique identifier                |
| MQTT Topic           | `#`             | Wildcard captures all sub-topics |
| Root Folder          | `IND360_folder` | Tag browser folder name          |
| JSON Payload         | ✅ Checked       | Auto-parses JSON into tags       |
| Create Writable Tags | ✅ Checked       | Enables writing back to device   |
| {% endstep %}        |                 |                                  |

{% step %}

### &#x20;Dashboard & Commands

Once tags appear in the Tag Browser under MQTT Engine, bind them to Vision components for live display.

<div align="center"><figure><img src="/files/XTkJUsvwx9CZr2j1Z06t" alt=""><figcaption></figcaption></figure></div>

#### Useful scripts

Below are the scripts that are used for publishing the data to the MQTT broker via buttons

**Tare Button Script**

```python
import json

# Build the tare command payload
payload = json.dumps({
    "$correlation": "IgnitionTare1",   # Unique ID to track this request
    "$ttl": 42,                        # Time-to-live in seconds
    "responseTopic": None,             # No response topic needed
    "command": "tare",                 # Command to execute on IND360
    "content": None                   # Not required for tare
})

# Publish to IND360 via Ignition MQTT Engine
system.cirruslink.engine.publish(
    "Ignition Distributor",                                     # Server name in MQTT Engine → Servers
    "ind360/devices/C249346600/commands/tare/invoke",           # Command topic
    payload.encode(),                                           # Must be bytes — .encode() converts string
    1,                                                          # QoS 1 = at least once delivery
    False                                                       # Retain = False
)
```

{% hint style="warning" %}
The payload must be byte-encoded using `.encode()`. Passing a plain string causes a silent failure; no error is shown, but the command is never published. Learn more about Python scripting here: <https://docs.chariot.io/display/CLD80/ME%3A+Python+Scripting>
{% endhint %}

**Clear Button Script**

```python
import json

# Build the tare-clear command payload
# Tare-clear resets the tare back to zero (removes the tare offset)
payload = json.dumps({
    "$correlation": "IgnitionClear1",  # Unique ID to track this request
    "$ttl": 42,                        # Time-to-live in seconds before message expires
    "responseTopic": None,             # Set to None if no response topic is needed
    "command": "tare-clear",           # The command to execute — clears the existing tare value
    "content": None                    # Additional content — not required for tare-clear command
})

# Publish the clear command to the MQTT broker via Ignition MQTT Engine
system.cirruslink.engine.publish(
    "Ignition Distributor",                                        # MQTT server name as configured in MQTT Engine > Servers
    "ind360/devices/C249346600/commands/tare-clear/invoke",        # MQTT topic — tare-clear command path for device C249346600
    payload.encode(),                                              # Payload must be bytes — .encode() converts string to bytes
    1,        
```

**Zero Button Script**

```python
import json

# Build the zero command payload as a Python dictionary
# Zero command sets the current scale reading as the new zero reference point
payload = json.dumps({
    "$correlation": "IgnitionZero1",   # Unique ID to track this request (can be any string)
    "$ttl": 42,                        # Time-to-live in seconds before message expires
    "responseTopic": None,             # Set to None if no response topic is needed
    "command": "zero",                 # The command to execute on IND360 (zero = set current weight as new zero)
    "content": None                    # Additional content — not required for zero command
})

# Publish the zero command to the MQTT broker via Ignition MQTT Engine
system.cirruslink.engine.publish(
    "Ignition Distributor",                                    # MQTT server name as configured in MQTT Engine > Servers
    "ind360/devices/C249346600/commands/zero/invoke",          # MQTT topic — device serial number C249346600 is part of the path
    payload.encode(),                                          # Payload must be bytes — .encode() converts string to bytes
    1,                                                         # QoS level 1 = at least once delivery (reliable)
    False                                                      # Retain flag — False means broker won't store this message
)
```

**Preset Tare Button Script**

* Take an input field and rename that to 'presetTare'
* Add the script below to the Preset Tare button.

```python
import json

# Read the preset value from the input field on the dashboard
preset_value = event.source.parent.getComponent("presetTare").value

if preset_value <= 0:
    system.gui.messageBox("Please enter a valid preset tare value", "Invalid Input")
else:
    payload = json.dumps({
        "$correlation": "IgnitionPresetTare1",
        "$ttl": 42,
        "responseTopic": None,
        "command": "tare-preset",
        "content": {
            "value": str(preset_value),     # Dynamic value from input field
            "unit": "Kg"
        }
    })
    system.cirruslink.engine.publish(
        "Ignition Distributor",
        "ind360/devices/C249346600/commands/tare-preset/invoke",
        payload.encode(), 1, False
    )
```

{% endstep %}

{% step %}

### Feedback Logging to Database

Log real device confirmation to SQLite using a Gateway Tag Change script triggered by the result timestamp.

**Create Database Table**

First, create a table in the default database. Run the following script in the Tools > Database Query browser.

```sql
CREATE TABLE IF NOT EXISTS ind360_command_log (
        id          INT AUTO_INCREMENT PRIMARY KEY,
        command     TEXT,
        result      TEXT,
        timestamp   TEXT
    )
```

To delete the table, use the following script

```sql
Drop table <table name>
```

**Gateway Tag Change Script:**

**tare\_result**

Watch tag: `[MQTT Engine]IND360_folder/ind360/devices/C249346600/commands/tare/result/_timestamp`

```python
def onTagChange(initialChange, newValue, previousValue, event, executionCount):
	
    # Get raw IND360 timestamp e.g. "2026-04-28T11:45:41.000Z"
    raw_timestamp = str(newValue.value)

    # Convert ISO 8601 format to SQLite compatible format
    # Replace the T separator with a space and remove the Z suffix
    clean_timestamp = raw_timestamp.replace("T", " ").replace("Z", "")
    # Result: "2026-04-28 11:45:41.000" ✅

    # Read result fields from MQTT Engine tags
    description = system.tag.readBlocking([
        "[MQTT Engine]IND360_folder/ind360/devices/C249346600/commands/tare/result/content/description"
    ])[0].value

    code = system.tag.readBlocking([
        "[MQTT Engine]IND360_folder/ind360/devices/C249346600/commands/tare/result/code"
    ])[0].value

    correlation = system.tag.readBlocking([
        "[MQTT Engine]IND360_folder/ind360/devices/C249346600/commands/tare/result/_correlation"
    ])[0].value

    # Log to database with cleaned SQLite compatible timestamp
    system.db.runPrepUpdate(
        "INSERT INTO ind360_command_log (command, result, timestamp) VALUES (?, ?, ?)",
        ["tare", description, clean_timestamp],
        database="Sample_SQLite_Database"
    )
```

**clear\_result**

Watch tag: `[MQTT Engine]IND360_folder/ind360/devices/C249346600/commands/clear/result/_timestamp`

```python
def onTagChange(initialChange, newValue, previousValue, event, executionCount):
	
    # Get raw IND360 timestamp e.g. "2026-04-28T11:45:41.000Z"
    raw_timestamp = str(newValue.value)

    # Convert ISO 8601 format to SQLite compatible format
    # Replace the T separator with a space and remove the Z suffix
    clean_timestamp = raw_timestamp.replace("T", " ").replace("Z", "")
    # Result: "2026-04-28 11:45:41.000" 

    # Read result fields from MQTT Engine tags
    description = system.tag.readBlocking([
        "[MQTT Engine]IND360_folder/ind360/devices/C249346600/commands/tare-clear/result/content/description"
    ])[0].value

    code = system.tag.readBlocking([
        "[MQTT Engine]IND360_folder/ind360/devices/C249346600/commands/tare-clear/result/code"
    ])[0].value

    correlation = system.tag.readBlocking([
        "[MQTT Engine]IND360_folder/ind360/devices/C249346600/commands/tare-clear/result/_correlation"
    ])[0].value

    # Log to database with cleaned SQLite compatible timestamp
    system.db.runPrepUpdate(
        "INSERT INTO ind360_command_log (command, result, timestamp) VALUES (?, ?, ?)",
        ["clear", description, clean_timestamp],
        database="Sample_SQLite_Database"
    )
```

**zero\_result**

Watch tag: `[MQTT Engine]IND360_folder/ind360/devices/C249346600/commands/zero/result/_timestamp`

```python
def onTagChange(initialChange, newValue, previousValue, event, executionCount):
	
    # Get raw IND360 timestamp e.g. "2026-04-28T11:45:41.000Z"
    raw_timestamp = str(newValue.value)

    # Convert ISO 8601 format to SQLite compatible format
    # Replace the T separator with a space and remove the Z suffix
    clean_timestamp = raw_timestamp.replace("T", " ").replace("Z", "")
    # Result: "2026-04-28 11:45:41.000" ✅

    # Read result fields from MQTT Engine tags
    description = system.tag.readBlocking([
        "[MQTT Engine]IND360_folder/ind360/devices/C249346600/commands/zero/result/content/description"
    ])[0].value

    code = system.tag.readBlocking([
        "[MQTT Engine]IND360_folder/ind360/devices/C249346600/commands/zero/result/code"
    ])[0].value

    correlation = system.tag.readBlocking([
        "[MQTT Engine]IND360_folder/ind360/devices/C249346600/commands/zero/result/_correlation"
    ])[0].value

    # Log to database with cleaned SQLite compatible timestamp
    system.db.runPrepUpdate(
        "INSERT INTO ind360_command_log (command, result, timestamp) VALUES (?, ?, ?)",
        ["zero", description, clean_timestamp],
        database="Sample_SQLite_Database"
    )
```

**preset\_tare**

Watch tag: `[MQTT Engine]IND360_folder/ind360/devices/C249346600/commands/tare-preset/result/_timestamp`

```python
def onTagChange(initialChange, newValue, previousValue, event, executionCount):
	
    # Get raw IND360 timestamp e.g. "2026-04-28T11:45:41.000Z"
    raw_timestamp = str(newValue.value)

    # Convert ISO 8601 format to SQLite compatible format
    # Replace the T separator with a space and remove the Z suffix
    clean_timestamp = raw_timestamp.replace("T", " ").replace("Z", "")
    # Result: "2026-04-28 11:45:41.000" ✅

    # Read result fields from MQTT Engine tags
    description = system.tag.readBlocking([
        "[MQTT Engine]IND360_folder/ind360/devices/C249346600/commands/tare-preset/result/content/description"
    ])[0].value

    code = system.tag.readBlocking([
        "[MQTT Engine]IND360_folder/ind360/devices/C249346600/commands/tare-preset/result/code"
    ])[0].value

    correlation = system.tag.readBlocking([
        "[MQTT Engine]IND360_folder/ind360/devices/C249346600/commands/tare-preset/result/_correlation"
    ])[0].value

    # Log to database with cleaned SQLite compatible timestamp
    system.db.runPrepUpdate(
        "INSERT INTO ind360_command_log (command, result, timestamp) VALUES (?, ?, ?)",
        ["tare", description, clean_timestamp],
        database="Sample_SQLite_Database"
    )
```

**Table SQL Query Binding**

To visualize the feedback in the table. Add a table to your screen and bind the data with the following script

```sql
SELECT
    command   AS Command,
    result    AS Result,
    timestamp AS Timestamp
FROM ind360_command_log
ORDER BY datetime(timestamp) DESC
LIMIT 50
```

**Clear button**

To clear the values in the table, you can use the following script

```python
# Show confirmation dialog before clearing
# Prevents accidental deletion of command history
result = system.gui.confirm(
    "Are you sure you want to clear all feedback records?",
    "Clear Feedback Table"
)

# Only delete if operator clicked Yes
if result:
    system.db.runUpdateQuery(
        "DELETE FROM ind360_command_log",
        database="Sample_SQLite_Database"
    )
```

{% endstep %}
{% endstepper %}

***

### Topic Structure Reference

```python
ind360/
├── devices/C249346600/
│ ├── scales/_/measurements #subscribe
│ ├── scales/_ #subscribe
│ ├── alarms #subscribe
│ └── commands/
│ ├── tare/invoke #publish
│ ├── tare/result #subscribe
│ ├── tare-clear/invoke #publish
│ ├── tare-clear/result #subscribe
│ ├── zero/invoke #publish
│ ├── zero/result #subscribe
│ ├── tare-preset/invoke #publish
│ └── tare-preset/result #subscribe
```

### Command Payload Reference

<table><thead><tr><th width="150.3333740234375">Command</th><th width="351.333251953125">content field</th><th>Notes</th></tr></thead><tbody><tr><td><code>tare</code></td><td><code>null</code></td><td>No content needed</td></tr><tr><td><code>tare-clear</code></td><td><code>null</code></td><td>No content needed</td></tr><tr><td><code>zero</code></td><td><code>null</code></td><td>No content needed</td></tr><tr><td><code>tare-preset</code></td><td><code>{"value": "2.0", "unit": "Kg"}</code></td><td>Value and unit required</td></tr></tbody></table>

#### Ignition Project

You can import this project in your Ignition designer to use it as a prototype:

{% file src="/files/87quKNpKqKT9Bz35QqCo" %}

***

## Key notes:

{% embed url="<https://www.canva.com/design/DAG7MP4dokE/MQMZPZt4YaA6um4lrJCE5A/view?utlId=hebd7857a29&utm_campaign=designshare&utm_content=DAG7MP4dokE&utm_medium=link2&utm_source=uniquelinks>" %}

***

## ♥️ Work With Me

I regularly test **industrial automation and IIoT devices**. If you’d like me to **review your product** or showcase it in my courses and YouTube channel:

📧 Email: <rajvir@codeandcompile.com> or drop me a message on [LinkedIn](https://www.linkedin.com/in/singhrajvir/)


---

# Agent Instructions: Querying This Documentation

If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://wiki.codeandcompile.com/product-reviews/smart-devices/smart-weighing-indicator/ind360-integration-with-mqtt.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.