> ## Documentation Index
> Fetch the complete documentation index at: https://docs.tryprofound.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Sentiment node

> Pull positive and negative sentiment scores from Profound into a Workflow

> [!TIP]
>
> To learn more about Sentiment metrics, see the [About Sentiment](https://docs.tryprofound.com/answer-engine-insights/sentiment/about.md) section.

The Sentiment node pulls current Sentiment data from your Profound account into a Workflow. It returns a positive or negative sentiment score for one or more brands or competitors, on a scale from 0 to 1. The score is model-balanced: it is the average across AI platforms, so no single platform controls the score.

Use the Sentiment node for tasks such as:

* Tracking positive and negative sentiment toward your brand over time
* Comparing sentiment between your brand and competitors
* Finding sentiment differences across platforms, regions, topics, or themes

> [!NOTE]
>
> The Sentiment node is the newer version of the [Sentiment Score node](https://docs.tryprofound.com/agents/workflows/nodes/data/sentiment-score.md), which uses an earlier version of Sentiment data.

## Node configuration

Selecting the Sentiment node opens its configuration panel on the right side of the Workflow Builder.

### Assets

Select one or more brands or competitors to query. The node queries each asset separately and labels each row with the asset name.

### Metric

Choose the sentiment metric to return:

* **Positive Sentiment** – The share of sentiment toward the asset that is positive.
* **Negative Sentiment** – The share of sentiment toward the asset that is negative.

The node returns each score as a ratio from 0 to 1. For example, `0.72` means 72%.

### Date range

Specify the time window to query. Options include:

* Last 2 Days
* Last 7 Days
* Last 14 Days
* Last 4 Weeks
* Last 3 Months
* Custom date range

If you select **Custom date range**, enter a start date and an end date.

### Filters

Use filters to narrow the dataset. Fields include:

* Tag
* Topics
* Platforms
* Personas
* Regions
* Theme

Select values from the list, or switch the filter to **Variable** to use a value from an earlier Workflow step. The Theme list comes from the first asset you selected.

> [!NOTE]
>
> When a filter value comes from a variable, Tag, Platforms, Personas, and Regions expect IDs. Topics and Theme accept names.

To group results by a filter field, open the filter's settings and select **Add \[field] as a dimension**. You can group by up to two dimensions.

### Output label

Examples:

* `sentiment`
* `weekly_positive_sentiment`
* `negative_sentiment_by_platform`

## Advanced settings

* **Date interval**: Group the results by day, week, or month.
* **Limit**: Set the maximum number of rows the node returns. The default is 100 and the maximum is 10,000. The limit applies to the combined results for all selected assets.
* **Sort by**: Sort the output by date or by the selected metric.
* **Sort direction**: Sort the output in ascending or descending order.

## Output

Returns a list of rows. Each row includes:

* `asset_name` – The brand or competitor.
* `date` – The date interval that the score covers.
* `metric_value` – The selected sentiment score, from 0 to 1.
* One field for each dimension you group by, such as `platform` or `region`.

This output is commonly used in:

* LLM narrative summaries
* Reputation monitoring Workflows
* Competitive sentiment reports

> [!TIP]
>
> Check out [Create a Workflow](https://docs.tryprofound.com/agents/workflows/create-workflow.md) to learn how to add this node to a Workflow.