> For the complete documentation index, see [llms.txt](https://docs.aisera.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.aisera.com/aisera-platform/tenant-setup/aisera-platform-configuration/tenant-configuration-settings/voice.md).

# Voice

Controls speech service routing, data retention, and Cisco Webex integration authentication for voice-enabled Aisera tenants.

The Voice configuration section contains tenant-level settings that govern how Aisera's voice capabilities operate. Settings here control audio recording retention, how the Aisera Platform authenticates inbound connections from Cisco Webex, and which speech service endpoints handle transcription and synthesis for the tenant.

### Audio recordings retention days

| **Type**    | Integer |
| ----------- | ------- |
| **Default** | `60`    |

Sets how many days the Aisera Platform retains voice call audio recordings before automatically deleting them from both the database and cloud object storage. Adjust this value to meet your organization's data governance or compliance requirements: lower it to reduce storage consumption, or raise it to extend access to historical recordings for quality review or audit purposes.

{% hint style="danger" %}
Deletion is permanent. You cannot recover recordings after the cleanup process deletes them. Setting this value to `0` causes all recordings to become immediately eligible for deletion on the next cleanup run.
{% endhint %}

## Cisco Webex integration

### Cisco webex organization ID

| **Type**    | Text field     |
| ----------- | -------------- |
| **Default** | Not configured |

Identifies your Cisco Webex organization so the Aisera Platform can authenticate inbound voice requests. When Webex sends a voice request, it includes a JSON Web Token, or JWT, with a `com.cisco.org.uuid` claim; the Aisera Platform compares that claim against this value and rejects any request where the claim does not match this value. Configure this field when setting up the Cisco Webex Bring Your Own Virtual Agent, or BYOVA, integration. Obtain your organization's universally unique identifier, or UUID, from your Webex administrator or the Webex Control Hub.

{% hint style="warning" %}
Although the UI marks this field as optional, the Cisco Webex integration requires it to function. Leaving it empty causes all inbound Cisco Webex voice requests to fail authentication.
{% endhint %}

See also: [Cisco webex data schema ID](#cisco-webex-data-schema-id)

### Cisco webex data schema ID

| **Type**    | Text field                             |
| ----------- | -------------------------------------- |
| **Default** | `5397013b-7920-4ffc-807c-e8a3e0a18f43` |

Identifies the Cisco Webex BYOVA data schema whose connections this tenant accepts. When Cisco Webex sends a voice connection, its JWT token contains a `com.cisco.datasource.schema.uuid` claim; Aisera rejects any connection where that claim does not exactly match this value. The default is Aisera's standard BYOVA schema UUID registered with Cisco, which is correct for all standard integrations. Override this only if your Cisco deployment uses a non-standard or tenant-specific schema UUID; your Cisco administrator or BYOVA setup documentation will provide the correct value.

{% hint style="warning" %}
This value must exactly match the `com.cisco.datasource.schema.uuid` claim in the Cisco-issued JWT token. Any mismatch, including differences in whitespace or capitalization, causes the Aisera Platform to reject all incoming Cisco voice connections.
{% endhint %}

See also: [Cisco webex organization ID](#cisco-webex-organization-id)

## Speech services

### Tenant Region

| **Type**    | Dropdown        |
| ----------- | --------------- |
| **Default** | `North America` |

Sets the geographic region for this tenant's voice services. The selected region determines the default endpoints used for Deepgram speech-to-text and ElevenLabs text-to-speech processing: North America routes to Deepgram's US cloud and ElevenLabs' standard endpoint; EMEA, or Europe, Middle East, and Africa, routes to Deepgram's EU deployment and ElevenLabs' EU residency endpoint. Set this to EMEA for any tenant whose voice traffic must remain within the EU to meet data residency or compliance requirements.

{% hint style="info" %}
Configuring the **Deepgram URL** or **ElevenLabs URL** fields overrides the region-based endpoint defaults. Ensure any manually configured endpoints match the intended region for your compliance requirements.
{% endhint %}

See also: [ElevenLabs URL](#elevenlabs-url)

### Deepgram URL

| **Type**    | Text field     |
| ----------- | -------------- |
| **Default** | Not configured |

Overrides the Deepgram speech-to-text endpoint for this tenant. If you leave this field empty, the service uses the regional default the **Tenant Region** setting determines (`wss://api.deepgram.com` for North America, `wss://api.eu.deepgram.com` for EMEA). Configure this when your deployment uses an on-premises or private Deepgram instance and you need to route a specific tenant's STT traffic to that endpoint.

{% hint style="info" %}
The URL must use WebSocket protocol: `wss://` for secure connections or `ws://` for insecure. For North America deployments, the text-to-speech, or TTS, path also derives its endpoint from this field by converting `wss://` to `https://`.
{% endhint %}

See also: [Tenant Region](#tenant-region)

### ElevenLabs URL

| **Type**    | Text field     |
| ----------- | -------------- |
| **Default** | Not configured |

Overrides the ElevenLabs text-to-speech endpoint for this tenant. If you leave this field empty, the service uses the regional default the **Tenant Region** setting determines (`https://api.elevenlabs.io/v1/` for North America, `https://api.eu.residency.elevenlabs.io/v1/` for EMEA). Configure this when your deployment uses a private or on-premises ElevenLabs instance and you need to route a specific tenant's TTS traffic to that endpoint.

{% hint style="info" %}
The URL must use HTTPS and include the trailing path prefix, for example: `https://your-instance.example.com/v1/` 
{% endhint %}

See also: [Tenant Region](#tenant-region)