> ## Documentation Index
> Fetch the complete documentation index at: https://axiom-mano-sample-app.mintlify.site/llms.txt
> Use this file to discover all available pages before exploring further.

# Configure Flow

> This page explains how to set up a flow to filter, shape, and route data from an Axiom dataset to a destination.

<Note>
  Flow is currently in private preview. To try it out, [sign up for a free preview](https://app.axiom.co/flows).
</Note>

A flow is a way to filter, shape, and route data from an Axiom dataset to a destination that you choose. This page explains how to set up a flow.

## Prerequisites

* [Create an Axiom account](https://app.axiom.co/register).
* [Create a dataset in Axiom](/reference/datasets#create-dataset) where you send your data.

{/* list separator */}

* Set up a destination. For more information, see [Destinations](/process-data/destinations).

## Set up a flow configuration

To set up a flow configuration:

1. Click the [Flows](https://app.axiom.co/flows) tab. Axiom displays the list of flow configurations you have created.

2. In the top right, click **New configuration**.

3. In the **Source** section, specify the source dataset and the transformation in an APL query. For example, the following APL query selects events from a `cloudflare-logpush` dataset and reduces them by removing a set of fields, before enriching with a new field.

   ```kusto
   ['cloudflare-logpush']
   | where QueryName == "app.axiom.co."
   // Reduce events by dropping unimportant field
   | project-away ['@app']*
   // Enrich events with additional context
   | extend ['@origin'] = "_axiom"
   ```

   <Note>
     If you only specify the name of the dataset in the query, Axiom routes all events to the destination.
   </Note>

4. Click **Preview** to check whether the query you specified transforms your data as desired. The **Input event** section displays the original data stored in Axiom. The **Output event** section displays the transformed data that Axiom sends to the destination. The original data in the Axiom dataset isn’t affected by the transformation.

5. In the **Destination** section, click **Add a destination**, and then select an existing destination where you want to route data or click **Create new destination**.

6. In the top right, click **Create**.

After creating a flow configuration, create a flow by selecting one of the following:

* **Continuous flow**
* **One-time flow**

## Create continuous flow

Continuous flows are continuously running operations that process your incoming data and route the outputs to a destination in real-time.

1. Click the **Flows** tab. Axiom displays the list of flow configurations you have created. Select the flow configuration that you want to use for creating a continuous flow.
2. In the top right, click **Create flow** and select **Continuous flow**.
3. Click **Create flow**.

As a result, Axiom starts running the query on all incoming data and routes the results of the query to the destination.

## Create one-time flow

One-time flows are one-off operations that process past data for a specific time range and route the output to a destination.

1. Click the **Flows** tab. Axiom displays the list of flow configurations you have created. Select the flow configuration that you want to use for creating a one-time flow.
2. In the top right, click **Create flow** and select **One-time flow**.
3. Specify the time range for events you want to process.
4. Click **Create flow**.

As a result, Axiom runs the query on the source data for the specified time range and routes the results of the query to the destination.

### Delivery rate of continuous flows

The delivery rate of a continuous flow currently depends on the rate at which you ingest data into the source dataset. For example, if you set up a continuous flow and ingest data at a rate of 1TB/day to the source dataset, Axiom processes events within a few seconds after ingest. If you ingest data to the source at a rate of 10GB/day, it can take several minutes for events to arrive. There is currently no maximum wait time for events to be processed by a continuous flow. As Flow progresses through the preview stage, Axiom will establish and refine maximum wait times.