Module voice 

The per-seat voice service: listeners, the sentence queue, and synthesis.

Re-exports

use std::collections::BTreeSet; 🔒

use std::collections::HashMap; 🔒

use std::sync::atomic::AtomicUsize; 🔒

use std::sync::atomic::Ordering; 🔒

use std::sync::Arc; 🔒

use std::sync::Mutex; 🔒

use forge_primitives::Message; 🔒

use forge_primitives::SessionSlot; 🔒

use forge_tts::Chunk; 🔒

use forge_tts::Engine; 🔒

use forge_tts::SpeakRequest; 🔒

use tokio::sync::mpsc; 🔒

use tokio::sync::oneshot; 🔒

use tracing::Instrument; 🔒

Structs

Job 🔒

PreviewSlot 🔒

Seat 🔒

VoiceService

The per-seat voice service. One render queue per seat, so a second say waits behind the first, and one engine for the whole process, loaded at boot and kept for the process’s life; a failed warm is retried by the first say.

VoiceSettings

What one utterance renders with, from the [tts] section.

WorkspaceNotes 🔒

The note dispatch a real workspace hands the service: the listener note enters the seat as a forge prompt, drawn as its own row.

WorkspaceSink 🔒

The sink a real workspace hands the service: every rendered clip crosses the update stream to the clients subscribed to the seat.

Enums

SayError

Why a say did not reach the sink.

Constants

LISTENING_NOTE

The note the first client’s turn-on puts into the seat.

NOT_LISTENING_NOTE

What the session is told when the last client leaves.

PREVIEW_LINE

The line a voice preview speaks, so two previews can be compared.

Traits

AudioSink

Where a rendered clip goes.

NoteDispatch

How a listener note reaches the seat’s session.

Functions

note_envelope 🔒

The envelope a note rides, so a view draws it as information rather than as the reader’s own turn.

render 🔒

Type Aliases

Loader 🔒