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

# Data & storage

> Choose the data source for the customer’s job. Airtable and Google Sheets support operational record work.

Choose the data source for the customer’s job. Airtable and Google Sheets support operational record work. A warehouse can provide a governed view of account and support context, with access and freshness maintained by your data team.

Runtime tools query a system when called. Knowledge ingestion prepares content for retrieval and has a separate connection, completion and refresh workflow; configuring one does not configure the other.

## Guides

- **[Retrieve customer context](/docs/guides/integrations/data-sources/retrieve-customer-context)** — authenticate the customer, read approved fields and handle missing or stale data.
- **[Record a support request](/docs/guides/integrations/data-sources/record-support-request)** — confirm details, save one intended request and reconcile uncertain creates.
- **[Update an existing record](/docs/guides/integrations/data-sources/update-existing-record)** — preserve record identity, apply an allowed change and read it back.

## Choose an integration

<a id="airtable" />
<a id="google-sheets" />

| Provider | Use the manual for |
| --- | --- |
| [Airtable](/docs/documentation/build/tools/airtable-tool) | Find records, save requests and update/remove authorized records using the built-in actions. |
| [Google Sheets](/docs/documentation/build/tools/google-sheets-tool) | Read ranges, append requests, update intended cells and create a team log using built-in actions. |
| [Snowflake](/docs/guides/integrations/snowflake/overview) | Read an approved customer view through the SQL API and a protected context service. |
| [Google BigQuery](/docs/guides/integrations/bigquery/overview) | Run a bounded, parameterized customer-context query with the service’s Google identity. |
| [Amazon Redshift](/docs/guides/integrations/redshift/overview) | Read customer context from Serverless through a completed Data API statement. |
| [Microsoft Fabric Warehouse](/docs/guides/integrations/microsoft-fabric/overview) | Read a Warehouse view through an encrypted SQL connection and Entra service principal. |
| [Databricks SQL](/docs/guides/integrations/databricks/overview) | Read a governed view on an AWS SQL warehouse using workspace OAuth and completed statement results. |

The warehouse recipes use a customer-hosted service to enforce the approved query, customer identity and source-age policy. Each manual includes its provider-specific permissions, executable setup, Voiceflow API-tool attachment and result checks. The data owner defines what the returned fields mean; a warehouse query does not refresh the upstream source.

## Build and maintain the workflow

<a id="developer-execution" />

Use [API tools](/docs/documentation/build/tools/api-tool) for a saved HTTP request, and [Run tools without a conversation](/docs/developers/tools/execute) for standalone schema and execution checks. Test agent routing and the customer’s answer separately after attachment/compilation.

<a id="knowledge-imports" />

For approved reference content, follow [Importing data sources](/docs/documentation/build/importing-data-sources). Verify what was selected, when ingestion completed and how changes are refreshed before using it in answers. Keep private transactional records within the authorized runtime workflow that owns them.

For another platform, follow [Create new integration](/docs/guides/integrations/create-new-integration).
