> For the complete documentation index, see [llms.txt](https://docs.kernel.ai/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.kernel.ai/integrations/write-source.md).

# Write source

Choose the connection Kernel writes data back to

Your **write source** is the connection selected to receive Kernel's write-back data. Kernel can read from multiple connected sources, while one connection is the write source for each environment.

In the app, the selected connection is labelled **Write source**. When you have multiple connections, **Connections → Write back to** lets you choose which one receives the data.

[Open Connections in Kernel](https://app.kernel.ai/settings/connection)

## Reading and writing

A connection gives Kernel access to a system or file exchange. Selecting it as the write source tells Kernel where to send supported changes and results. Other connected sources remain available for reading.

For example, you might connect both Salesforce and Amazon S3, with Salesforce selected as the write source. Kernel can read the connected source data, and account actions such as **Create** use Salesforce as their destination.

The write source and field mappings work together:

* **Write source** chooses the destination connection.
* **Field mappings or output configuration** choose the fields and values to write.
* **The action, sync or export** determines which records are sent and when.

Sandbox and production have separate write-source selections. Check the environment when reviewing the destination.

## Available write sources

| Write source   | How Kernel writes data                                                                                                                                                                                                        |
| -------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| **Salesforce** | Writes mapped data to Salesforce records and supports account actions such as Create, Associate, Merge and Delete. The connected user's permissions and your configuration determine which changes can be made.               |
| **Amazon S3**  | Delivers results as CSV files to the configured output bucket and location. Your team can use these files in its downstream systems. Configure a write role and output bucket before selecting Amazon S3 as the write source. |

Each integration has its own workflows. See [Salesforce integration](/integrations/salesforce-integration.md) for connection and permissions, or [S3 file exchange](/integrations/s3-integration.md) for file input, output and delivery settings.

## Choose or change the write source

1. Open **Connections** and select the intended environment.
2. Under **Write back to**, find the connection you want to use and select **Make write source**.
3. Review the source and environment in the confirmation.
4. Type the source name shown, such as **Salesforce** or **Amazon S3**, then select **Change write source**.
5. Check that the connection is labelled **Write source**.

There is only one write source at a time in each environment. Changing it replaces the previous selection. The previous connection remains available for reading, and write-back workflows use the newly selected connection.

Changing the selection does not move existing records or start a data transfer. Review the destination's mappings or output configuration, then run the relevant action, sync or export when ready.

## How this affects Create accounts

[Create accounts](/platform/actioning/create.md) adds reviewed Kernel accounts to your write source. The **New accounts** field mappings determine the values used for each new record, and Kernel links the result to the existing Kernel account.

The app names the destination in the confirmation and result. With Salesforce selected, for example, an existing-record result is labelled **Already in Salesforce**.

**Create accounts currently supports Salesforce as the write source.** For Amazon S3, use the file exchange workflow to deliver account data in CSV files.

## FAQs and troubleshooting

<details>

<summary>Why can I not see Write back to?</summary>

**Write back to** appears when multiple connections are available. With one connection, check its **Write source** label on **Connections**. Also confirm that you are viewing the intended environment.

</details>

<details>

<summary>Why can I not select Amazon S3 as the write source?</summary>

Configure both a write role and an output bucket for that connection. A connection set up only for reading needs these output settings before it can be selected. See [S3 output configuration](/integrations/s3-integration.md#output-configuration) for the setup details.

</details>

<details>

<summary>Does selecting a write source send data immediately?</summary>

The selection sets the destination. Data is sent when the relevant action, sync or export runs. Check that workflow's result to confirm which records were written.

</details>

<details>

<summary>Where should I check if data has not appeared?</summary>

Check the **Write source** label, environment, connection permissions and saved mappings or output configuration. Then review the result of the action, sync or export. For Salesforce, inspect the destination record and mapped fields. For Amazon S3, check the configured output location and export result.

</details>

## Related pages

* [Create accounts](/platform/actioning/create.md)
* [Salesforce integration](/integrations/salesforce-integration.md)
* [S3 file exchange](/integrations/s3-integration.md)