Cognigy Voice Gateway¶
Overview¶

Within our Cognigy.AI platform you are able to connect your Agent to your Contact Center or other phone numbers using our VoiceGateway, allowing customers to speak with your Agent instead of just writing to it.
Voice Gateway Specific Nodes¶
Cognigy.AI comes with built-in Nodes to control Voice Gateway. See Voice Gateway Nodes for more information.
Call Meta Data¶
Voice Gateway identifies information about the caller and adds it to the Cognigy Input Object as input.data.numberMetaData
.
Parameter | Type | Description | Example |
---|---|---|---|
headers | JSON | The SIP Headers of the call on INVITE, including Custom Headers | See example below |
number | string | The phone number of the caller, including country code | +4921154591991 |
country | string | The 2-character country code | DE |
countryCallingCode | string | The calling code of the country | 49 |
nationalNumber | string | The national number without the country code and without a leading zero. | 21154591991 |
valid | boolean | Whether the number is valid | true |
valid | string | The type of number. See below. | FIXED_LINE |
uri | string | The URI for the number | tel:+4921154591991 |
numberMetaData.type
can be any of:
- PREMIUM_RATE
- TOLL_FREE
- SHARED_COST
- VOIP
- PERSONAL_NUMBER
- PAGER
- UAN
- VOICEMAIL
- FIXED_LINE_OR_MOBILE
- FIXED_LINE
- MOBILE
{
"from": "<caller-phone-number>",
"to": "<sip-destination>",
"call-id": "<id-value>",
"allow": "NOTIFY, OPTIONS, BYE, INVITE, ACK, CANCEL, REFER",
"X-Custom-Headers": "<custom-headers-value>",
"X-Originating-Carrier": "<carrier-name>",
"X-Voip-Carrier-Sid": "<id-value>",
"X-Twilio-AccountSid": "<id-value>",
"X-Twilio-CallSid": "<id-value>",
"other-properties": "..."
}
NumberMetaData in Tokens
All of the above are available as Tokens inside Cognigy Text fields as well.
Call Events¶
Allows activating call events for a Flow. Select a call event from the Voice Gateway Events list. This event that will trigger the action.
If you have configured the same call event in both the Endpoint and the Lookup Node, the Endpoint settings will overwrite the Node settings.
Call Event Settings¶
Note
As with all other Endpoint settings, you cannot test the Call Events settings in the Endpoint within the Interaction Panel.
Parameter | Type | Description |
---|---|---|
Action | Selector | Choose the action to be performed when the call event is detected: - Inject into current Flow — inject the defined text and data payload into the current flow. - Execute Flow — trigger a selected Flow when the call event is detected. - None — no action will be taken when the call event is detected. - Transfer — transfer a call in case TRANSFER_DIAL_ERROR or TRANSFER_REFER_ERROR call events occur. This action for call events is configured similarly to Call Failover. |
Text Payload | String | Enter the text that will be sent to your Flow. Available only for the Inject into current Flow action. |
Data Payload | JSON | Provide the data that will be sent into your Flow in JSON format. Available only for the Inject into current Flow action. |
Execute Flow | Selector | Execute the selected Flow. Flow execution will stop afterward. Available only for the Execute Flow action. |
Call Failover¶
The Call Failover section is intended to handle runtime errors that occur on the Cognigy.AI side during the execution of a Flow.
Feature availability
- If you use a SaaS Cognigy installation, contact the support team to activate this feature.
- If you use an On-Premise Cognigy installation, activate this feature by adding
FEATURE_ENABLE_ENDPOINT_CALL_FAILOVER
invalues.yaml
.
Parameter | Type | Description | Transfer Type |
---|---|---|---|
Failover enabled | Toggle | If enabled, the configuration below will be used to perform a transfer in case of a runtime error | - |
Transfer Type | Dropdown | There are two transfer types: - Refer — forwarding an existing call. - Dial — creating a new outgoing call. If you want to use this type and still have the old Node version, add a new Voice Gateway Transfer Node in the Flow Editor and manually transfer the required settings from the old Node. |
- |
Reason | CognigyScript | The reason for the handover. It is shown in Voice Gateway logs. | All |
Target | CognigyScript | E.164 syntax or a SIP URI are supported. | All |
Caller ID | Number | The caller ID. Some carriers, like Twilio, require a registered number for outgoing calls. | Dial |
Dial Music | URL | Custom audio or ring-back which plays to the caller while the outbound call is ringing. Only the .wav or .mp3 formats are supported. |
Dial |
Timeout | Number | The amount of time (in seconds) that the virtual agent will ring before a no-answer timeout. The default value is 60 seconds. | Dial |
Failover Transcribe Enabled | Toggle | If enabled, transcriptions will be attempted in case of a failed call transfer. | Dial |
STT Vendor | Selector | Select the desired STT Vendor. For custom use all lowercase letters like microsoft, google, aws, nuance, or deepgram. | Dial |
STT Language | Selector | Select the desired STT Language. For custom languages, use the following format: de-DE, fr-FR, en-US. | Dial |
Disable STT Punctuation | Toggle | This parameter is active only when Google or Deepgram is selected in the STT Vendor setting. Prevents the STT response from the virtual agent to include punctuation marks. |
Dial |
Deepgram Tier | Dropdown | This parameter is active only when Deepgram is selected in the STT Vendor setting. Choose a tier for your API request, and ensure that the model is available for the chosen STT language. For detailed information about Deepgram tiers, refer to the Deepgram documentation. |
Dial |
Deepgram Model | Dropdown | This parameter is active only when Deepgram is selected in the STT Vendor setting. Choose a model for processing submitted audio. Each model is associated with a tier. Ensure that the selected tier is available for the chosen STT language. For detailed information about Deepgram models, refer to the Deepgram documentation. |
Dial |
Endpointing | Toggle | This parameter is active only when Deepgram is selected in the STT Vendor setting. Deepgram's Endpointing feature watches streaming audio for long pauses that signal the end of speech. When it spots an endpoint, it finalizes predictions and returns the transcript, marking it as complete with the speech_final parameter set to true . For detailed information about Deepgram Endpointing, refer to the Deepgram documentation.The duration for detecting the end of speech is preconfigured with a default value (10 milliseconds). If you want to change this value, use the Endpointing Time setting. |
Dial |
Endpointing Time | Number | This parameter is active only when Deepgram is selected in the STT Vendor setting and the Endpointing toggle is enabled. Customize the duration (in milliseconds) for detecting the end of speech. The default is 10 milliseconds of silence. Transcripts are sent after detecting silence, and the system waits until the speaker resumes or the required silence time is reached. Once either condition is met, a transcript is sent back with speech_final set to true . |
Dial |
Smart Formatting | Toggle | This parameter is active only when Deepgram is selected in the STT Vendor setting. Deepgram's Smart Format feature applies additional formatting to transcripts to optimize them for human readability. Smart Format capabilities vary between models. When Smart Formatting is turned on, Deepgram will always apply the best-available formatting for your chosen model, tier, and language combination. For detailed examples, refer to the Deepgram documentation. Note that when Smart Formatting is turned on, punctuation will be activated, even if you have the Disable STT Punctuation setting enabled. |
Dial |
Google Model | Dropdown | This parameter is active only when Google is selected in the STT Vendor setting. Utilizes one of Google Cloud Speech-to-Text transcription models, with the latest_short model being the default choice. For a detailed list of Google models, refer to the Transcription models section in the Google Documentation. Keep in mind that the default value is a Google Model type that can be used if other models don't suit your specific scenario. |
Dial |
Transcription Webhook | URL | The Webhook is triggered with an HTTP POST whenever an interim or final transcription is received. Uses the default recognizer. | Dial |
STT Label | CognigyScript | The alternative name of the vendor is the one you specify in the Voice Gateway Self-Service Portal. If you have created multiple speech services from the same vendor, use the label to specify which service to use. | Dial |
Audio Stream Selection | Selector | Select the source of the audio stream: - Caller/Called - both the incoming and outgoing audio streams of the caller and the called party. - Caller - the incoming and outgoing audio stream of the caller. - Called - the incoming and outgoing audio stream of the called party. Ensure that the selected audio stream matches the language specified for transcription. If no audio stream is provided, the system will use the one set in the beginning, which should also match the language specified for transcription. |
Dial |
Custom Transfer SIP Headers | Toggle | Data that needs to be sent as SIP headers in the generated SIP message. | All |
Generic Endpoint Settings¶
Find out about the generic Endpoint settings available with this Endpoint on the following pages:
- Endpoints Overview
- Data Protection & Analytics
- Transformer Functions
- NLU Connectors
- Session Management
- Real Time Translation Settings
Contact Center & Phone number linking
In order to route your Contact Center or Phone Number to your Voice Gateway Endpoint, get in touch with us via an email to support@cognigy.com.
Rebranding of Voice Gateway with AudioCodes
With the native Voice Gateway integration to Cognigy AI the AudioCodes implementation will be rebranded from Voice Gateway to AudioCodes. This applies to the Flow Nodes and the Endpoint.