Publish changes to a deployed voice agent

PUT

/voice/agents/{agentId}

const url = 'https://api.pingram.io/voice/agents/example';
const options = {
  method: 'PUT',
  headers: {Authorization: 'Bearer <token>', 'Content-Type': 'application/json'},
  body: '{"spec":{"name":"example","instructions":"example","inbound":{"firstAction":"speak","greeting":"example"},"outbound":{"firstAction":"speak","opener":"example","voicemailAction":"hangup","voicemailMessage":"example"},"model":{"mode":"chained","llm":"example","stt":"example","tts":"example","voiceId":"example","language":"example","speechSpeed":1,"temperature":1,"maxTokens":1},"tools":[{"type":"transfer","transferMode":"warm","destination":"example","briefing":"example"}],"variables":[{"name":"example","description":"example","defaultValue":"example"}],"conversation":{"turnDetection":"semantic","minEndOfTurnSilenceMs":1,"allowInterruptions":true,"minInterruptionDurationMs":1,"silenceTimeoutSeconds":1,"maxCallLengthSeconds":1,"agentCanEndCall":true},"compliance":{"recordingEnabled":true}}}'
};

try {
  const response = await fetch(url, options);
  const data = await response.json();
  console.log(data);
} catch (error) {
  console.error(error);
}

curl --request PUT \
  --url https://api.pingram.io/voice/agents/example \
  --header 'Authorization: Bearer <token>' \
  --header 'Content-Type: application/json' \
  --data '{ "spec": { "name": "example", "instructions": "example", "inbound": { "firstAction": "speak", "greeting": "example" }, "outbound": { "firstAction": "speak", "opener": "example", "voicemailAction": "hangup", "voicemailMessage": "example" }, "model": { "mode": "chained", "llm": "example", "stt": "example", "tts": "example", "voiceId": "example", "language": "example", "speechSpeed": 1, "temperature": 1, "maxTokens": 1 }, "tools": [ { "type": "transfer", "transferMode": "warm", "destination": "example", "briefing": "example" } ], "variables": [ { "name": "example", "description": "example", "defaultValue": "example" } ], "conversation": { "turnDetection": "semantic", "minEndOfTurnSilenceMs": 1, "allowInterruptions": true, "minInterruptionDurationMs": 1, "silenceTimeoutSeconds": 1, "maxCallLengthSeconds": 1, "agentCanEndCall": true }, "compliance": { "recordingEnabled": true } } }'

Authorizations

Parameters

Path Parameters

agentId

required

string

Agent id

Request Body^required

application/json

object

spec

required

object

name

required

string

instructions

required

string

inbound

required

object

firstAction

required

string

Allowed values: speak wait

greeting

required

string

outbound

required

object

firstAction

required

string

Allowed values: speak wait

opener

required

string

voicemailAction

required

string

Allowed values: hangup message continue

voicemailMessage

string

model

required

One of:

Chained
S2s

Chained

Separate STT, LLM, and TTS providers. For best call quality, prefer s2s mode.

object

mode

required

string

Allowed values: chained

llm

required

‘provider:model’, e.g. ‘openai:gpt-4o’

string

stt

required

‘provider:model’, e.g. ‘deepgram:nova-3’

string

tts

required

‘provider:model’, e.g. ‘elevenlabs:eleven_multilingual_v2’

string

voiceId

required

Provider-native voice id for the selected TTS provider (e.g. ElevenLabs UUID, OpenAI alloy).

string

language

required

string

speechSpeed

required

number

temperature

required

number

maxTokens

required

number

S2s

Speech-to-speech realtime model — best call quality.

object

mode

required

string

Allowed values: s2s

model

required

‘provider:model’, e.g. ‘openai:gpt-realtime’

string

voiceId

required

Provider-native voice id for the selected realtime model (e.g. OpenAI marin).

string

temperature

required

number

maxTokens

required

number

tools

required

Array

One of:

Transfer

object

type

required

string

Allowed values: transfer

transferMode

required

string

Allowed values: warm cold

destination

required

string

briefing

string

variables

required

Array<object>

object

name

required

string

description

string

defaultValue

string

conversation

required

object

turnDetection

required

string

Allowed values: semantic vad

minEndOfTurnSilenceMs

required

number

allowInterruptions

required

boolean

minInterruptionDurationMs

required

number

silenceTimeoutSeconds

required

number

maxCallLengthSeconds

required

number

agentCanEndCall

When true (default), the agent may invoke the built-in end_call action.

boolean

compliance

required

object

recordingEnabled

required

boolean

Responses

200

Successful response

application/json

object

agent

required

Agent record returned by GET /voice/agents.

object

agentId

required

string

deployment

required

string

Allowed values: hosted ejected

spec

required

object

name

required

string

instructions

required

string

inbound

required

object

firstAction

required

string

Allowed values: speak wait

greeting

required

string

outbound

required

object

firstAction

required

string

Allowed values: speak wait

opener

required

string

voicemailAction

required

string

Allowed values: hangup message continue

voicemailMessage

string

model

required

One of:

Chained
S2s

Chained

Separate STT, LLM, and TTS providers. For best call quality, prefer s2s mode.

object

mode

required

string

Allowed values: chained

llm

required

‘provider:model’, e.g. ‘openai:gpt-4o’

string

stt

required

‘provider:model’, e.g. ‘deepgram:nova-3’

string

tts

required

‘provider:model’, e.g. ‘elevenlabs:eleven_multilingual_v2’

string

voiceId

required

Provider-native voice id for the selected TTS provider (e.g. ElevenLabs UUID, OpenAI alloy).

string

language

required

string

speechSpeed

required

number

temperature

required

number

maxTokens

required

number

S2s

Speech-to-speech realtime model — best call quality.

object

mode

required

string

Allowed values: s2s

model

required

‘provider:model’, e.g. ‘openai:gpt-realtime’

string

voiceId

required

Provider-native voice id for the selected realtime model (e.g. OpenAI marin).

string

temperature

required

number

maxTokens

required

number

tools

required

Array

One of:

Transfer

object

type

required

string

Allowed values: transfer

transferMode

required

string

Allowed values: warm cold

destination

required

string

briefing

string

variables

required

Array<object>

object

name

required

string

description

string

defaultValue

string

conversation

required

object

turnDetection

required

string

Allowed values: semantic vad

minEndOfTurnSilenceMs

required

number

allowInterruptions

required

boolean

minInterruptionDurationMs

required

number

silenceTimeoutSeconds

required

number

maxCallLengthSeconds

required

number

agentCanEndCall

When true (default), the agent may invoke the built-in end_call action.

boolean

compliance

required

object

recordingEnabled

required

boolean

phoneNumbers

required

Array<string>

workerPresence

string

Allowed values: online offline

createdAt

required

string

updatedAt

required

string

Example

{
  "agent": {
    "deployment": "hosted",
    "spec": {
      "inbound": {
        "firstAction": "speak"
      },
      "outbound": {
        "firstAction": "speak",
        "voicemailAction": "hangup"
      },
      "model": {
        "mode": "chained"
      },
      "tools": [
        {
          "type": "transfer",
          "transferMode": "warm"
        }
      ],
      "conversation": {
        "turnDetection": "semantic"
      }
    },
    "workerPresence": "online"
  }
}

400

Bad Request - validation errors, invalid input

application/json

Standard error response for API errors.

object

trackingId

required

Unique tracking ID for the request.

string

error

required

Structured error details for API error responses.

object

code

required

Machine-readable error code.

string

message

required

Human-readable error message.

string

fix

Actionable hint for fixing the error.

string

Example^generated

{
  "trackingId": "example",
  "error": {
    "code": "example",
    "message": "example",
    "fix": "example"
  }
}

401

Unauthorized

402

Payment Required - usage limits exceeded

application/json

Standard error response for API errors.

object

trackingId

required

Unique tracking ID for the request.

string

error

required

Structured error details for API error responses.

object

code

required

Machine-readable error code.

string

message

required

Human-readable error message.

string

fix

Actionable hint for fixing the error.

string

Example^generated

{
  "trackingId": "example",
  "error": {
    "code": "example",
    "message": "example",
    "fix": "example"
  }
}

500

Internal Server Error

502

Bad Gateway - provider error

application/json

Standard error response for API errors.

object

trackingId

required

Unique tracking ID for the request.

string

error

required

Structured error details for API error responses.

object

code

required

Machine-readable error code.

string

message

required

Human-readable error message.

string

fix

Actionable hint for fixing the error.

string

Example^generated

{
  "trackingId": "example",
  "error": {
    "code": "example",
    "message": "example",
    "fix": "example"
  }
}

Publish changes to a deployed voice agent

Authorizations

Parameters

Path Parameters

Request Bodyrequired

Responses

200

Example

400

Examplegenerated

401

402

Examplegenerated

500

502

Examplegenerated

Request Body^required

Example^generated

Example^generated

Example^generated