dial_sdk.self_hosted_audio¶

Self-Hosted AUDIO protocol — the WebSocket contract your server speaks when Dial pipes the raw call audio to you, full duplex (https://docs.getdial.ai/api-reference/self-hosted-audio-protocol).

These pydantic models are the executable form of that contract, plus parse_dial_audio_message / serialize_server_audio_message. Verify the connection with verify_dial_signature (from dial_sdk.self_hosted). Configuration of Self-Hosted mode happens in the dashboard / REST API — this module is purely the protocol types for building a conformant server; you own the WebSocket loop (see the audio echo playbook).

Attributes¶

`AudioFormat`
`CallEndReason`
`DialAudioMessage`
`ServerAudioMessage`

Classes¶

`AudioCallConnected`
`AudioCallEnded`
`AudioDtmf`
`AudioDurationWarning`
`AudioFormats`
`AudioMark`
`AudioMedia`
`ServerAudioMedia`
`ServerClear`
`ServerEndCall`
`ServerMark`

Functions¶

`parse_dial_audio_message`(→ DialAudioMessage)	Parse + validate a frame received from Dial. Raises `ValidationError`.
`serialize_server_audio_message`(→ str)	Validate + serialize an outbound frame to a JSON string.

Module Contents¶

class dial_sdk.self_hosted_audio.AudioCallConnected¶

Bases: pydantic.BaseModel

call_id: str¶

direction: Literal['inbound', 'outbound']¶

formats: AudioFormats¶

from_: str¶

instruction: str | None = None¶

language: str | None = None¶

model_config¶

reconnect: bool = False¶

to: str¶

type: Literal['call_connected']¶

class dial_sdk.self_hosted_audio.AudioCallEnded¶

Bases: pydantic.BaseModel

reason: CallEndReason¶

type: Literal['call_ended']¶

class dial_sdk.self_hosted_audio.AudioDtmf¶

Bases: pydantic.BaseModel

digit: str¶

type: Literal['dtmf']¶

class dial_sdk.self_hosted_audio.AudioDurationWarning¶

Bases: pydantic.BaseModel

seconds_remaining: int¶

type: Literal['duration_warning']¶

class dial_sdk.self_hosted_audio.AudioFormats¶

Bases: pydantic.BaseModel

inbound: AudioFormat¶

outbound: AudioFormat¶

class dial_sdk.self_hosted_audio.AudioMark¶

Bases: pydantic.BaseModel

name: str¶

type: Literal['mark']¶

class dial_sdk.self_hosted_audio.AudioMedia¶

Bases: pydantic.BaseModel

payload: str¶

seq: int | None = None¶

type: Literal['media']¶

class dial_sdk.self_hosted_audio.ServerAudioMedia¶

Bases: pydantic.BaseModel

payload: str¶

type: Literal['media'] = 'media'¶

class dial_sdk.self_hosted_audio.ServerClear¶

Bases: pydantic.BaseModel

type: Literal['clear'] = 'clear'¶

class dial_sdk.self_hosted_audio.ServerEndCall¶

Bases: pydantic.BaseModel

type: Literal['end_call'] = 'end_call'¶

class dial_sdk.self_hosted_audio.ServerMark¶

Bases: pydantic.BaseModel

name: str¶

type: Literal['mark'] = 'mark'¶

dial_sdk.self_hosted_audio.parse_dial_audio_message(raw: str | bytes | dict) → DialAudioMessage¶: Parse + validate a frame received from Dial. Raises ValidationError.

dial_sdk.self_hosted_audio.serialize_server_audio_message(message: ServerAudioMessage) → str¶: Validate + serialize an outbound frame to a JSON string.

dial_sdk.self_hosted_audio.AudioFormat¶

dial_sdk.self_hosted_audio.CallEndReason¶

dial_sdk.self_hosted_audio.DialAudioMessage¶

dial_sdk.self_hosted_audio.ServerAudioMessage¶