Skip to content
Voice/
API Reference
/
/
SVAML Instructions
Last updated

Instructions

Instructions allow your application to perform tasks during a call, such as play a message to participants given a particular locale, or starting a recording. The following instructions are available to use:

playFiles

Plays Interactive Voice Response (IVR) files for the supported locale or SSML commands at the Sinch backend. An IVR message is played only on the caller's side.

Available to use in a response to an Incoming Call Event callback, an Answered Call Event callback, or a Prompt Input Event callback.

Example code

application/json

The playFiles instruction.

{
  "instructions": [
    { … }
  ]
}

Schema

namestringrequired

The name property. Must have the value playFiles.

Default "playFiles"
ValueDescription
playFiles

The playFiles instruction.

idsArray of stringsrequired

The IDs of the files which will be played. These can be a URL to a file, SSML commands using the #ssml[] element, or text using the #tts[] element.

Example: ["#ssml[Thank you for calling Sinch!]"]
localestring

If using SSML or TTS, this is a required field. The voice and language you want to use for the text-to-speech message. This can either be defined by the ISO 639 locale and language code or by specifying a particular voice. Supported languages and voices are detailed here.

Example: "en-US"

say

Plays a synthesized text-to-speech message to the end user. The message is provided in the text field.

Available to use in a response to an Incoming Call Event callback, an Answered Call Event callback, or a Prompt Input Event callback.

Example code

application/json

The say instruction.

{
  "instructions": [
    { … }
  ]
}

Schema

namestringrequired

The name property. Must have the value say.

Default "say"
ValueDescription
say

The say instruction.

textstring

Contains the message that will be spoken. Default maximum length is 600 characters. To change this limit, please contact support.

Example: "Hello, thank you for calling Sinch!"
localestring

The voice and language you want to use for the text-to-speech message. This can either be defined by the ISO 639 locale and language code or by specifying a particular voice. Supported languages and voices are detailed here.

Example: "en-US"

sendDtmf

Plays DTMF tones in the call.

Available to use in a response to an Incoming Call Event callback or a Prompt Input Event callback.

Example code

application/json

The sendDtmf instruction.

{
  "instructions": [
    { … }
  ]
}

Schema

namestringrequired

The name property. Must have the value sendDtmf.

Default "sendDtmf"
ValueDescription
sendDtmf

The sendDtmf instruction.

valuestring

A string that determines the DTMF tones to play to the callee when the call is picked up. Valid characters are: 0-9, #, and w. w renders a 500ms pause. For example, the string ww1234#w#, plays a 1 second pause, the DTMF tones for 1, 2, 3, 4, and #, followed by a 500ms pause and finally the # tone. This is useful if the callout destination requires a conference PIN code or an extension. If there is a calling party, it will hear progress while the DTMF is sent.

Example: "ww1234#w#"

setCookie

Creates a cookie for the call.

Available to use in a response to an Incoming Call Event callback, an Answered Call Event callback, or a Prompt Input Event callback.

Example code

application/json

The setCookie instruction.

{
  "instructions": [
    { … }
  ]
}

Schema

namestringrequired

The name property. Must have the value setCookie.

Default "setCookie"
ValueDescription
setCookie

The setCookie instruction.

keystring

The name of the cookie you want to set.

Example: "cookie_name"
valuestring

The value of the cookie you want to set.

Example: "cookie_value"

If you set a cookie in the response to an event (like “ice” or “ace”), it will be sent along in subsequent events in the session (like for instance “dice”). This can be used to keep simple application state during a call.

Limitations

The total size of cookie data (keys and values) may not exceed 1024 bytes.

answer

Forces the callee to answer the call.

Example code

application/json

The answer instruction.

{
  "instructions": [
    { … }
  ]
}

Schema

namestringrequired

The name property. Must have the value answer.

Default "answer"
ValueDescription
answer

The answer instruction.

Normally, a call is “answered” when a callee picks up a call from a caller (for instance, when the callee picks up when doing a connectPstn action for an incoming call, see connectPstn). This normally means that the call duration starts “ticking” on the caller's phone and that sound can flow in both directions.

This is normally handled automatically, but by adding the answer instruction, you can force the incoming call to be answered before the call is connected and a callee picks up. For instance, if you add the answer instruction in an ICE response with the connectPstn action, the caller's phone will pick up and its call duration will start ticking immediately (and not wait until the callee answers the call).

startRecording

Starts recording the call.

Available to use in a response to an Incoming Call Event callback, an Answered Call Event callback, or a Prompt Input Event callback.

Example code

application/json

The startRecording instruction.

{
  "instructions": [
    { … }
  ]
}

Schema

namestringrequired

The name property. Must have the value startRecording.

Default "startRecording"
ValueDescription
startRecording

The startRecording instruction.

optionsobject(svaml.instruction.startRecordingOptions)
Note:

Recording calls when using the connectMxp action is not supported.

stopRecording

Stops an ongoing recording for the current call.

Available to use in a response to a Prompt Input Event callback.

Example code

application/json

The stopRecording instruction.

{
  "instructions": [
    { … }
  ]
}

Schema

namestringrequired

The name property. Must have the value stopRecording.

Default "stopRecording"
ValueDescription
stopRecording

The stopRecording instruction.

Previous page
Next page