Add live WebRTC voice conversations to the TUI (#43581)

## What changed

- Add feature-gated `/voice`, `/voice mute`, and `/voice stop` commands with local WebRTC audio and app-server signaling.
- Show live transcripts, conversation status, and microphone and speaker levels. Preserve captions and undelivered answers across thread switches.
- Speak final answers from voice handoffs while keeping delegated reasoning and commentary hidden and typed answers unspoken.
- Retry eligible startup failures once and clean up voice sessions on thread switches and disconnects.
- Stop voice and block late handoffs after a misalignment policy violation. Remove realtime event payloads and spoken text from receipt and debug logs.

## Testing

Add coverage for voice command mapping, rejected starts, disconnect cleanup, delegated final-answer speech, late handoff rejection, and transcript and footer rendering.

GitOrigin-RevId: 11def1f86b0e023b21d0d92072d6636ccad89815

codex-rs/tui/src/app_command.rs

@@ -1,3 +1,4 @@
+use std::fmt;
 use std::path::PathBuf;
 
 use codex_app_server_protocol::AskForApproval;
@@ -9,6 +10,7 @@ use codex_app_server_protocol::ReviewTarget;
 use codex_app_server_protocol::ToolRequestUserInputResponse;
 use codex_app_server_protocol::UserInput;
 use codex_config::types::ApprovalsReviewer;
+use codex_protocol::ThreadId;
 use codex_protocol::approvals::GuardianAssessmentEvent;
 use codex_protocol::config_types::CollaborationMode;
 use codex_protocol::config_types::Personality;
@@ -19,13 +21,98 @@ use codex_protocol::models::PermissionProfile;
 use codex_protocol::openai_models::ReasoningEffort as ReasoningEffortConfig;
 use codex_protocol::request_permissions::RequestPermissionsResponse;
 use serde::Serialize;
+use serde::Serializer;
 use serde_json::Value;
 
+/// Keeps ICE credentials out of command diagnostics and session recordings.
+/// Convert back to a string only when sending the offer to app-server signaling.
+#[derive(Clone, PartialEq, Eq)]
+pub(crate) struct RealtimeOfferSdp(String);
+
+impl From<String> for RealtimeOfferSdp {
+    fn from(value: String) -> Self {
+        Self(value)
+    }
+}
+
+impl From<RealtimeOfferSdp> for String {
+    fn from(value: RealtimeOfferSdp) -> Self {
+        value.0
+    }
+}
+
+impl fmt::Debug for RealtimeOfferSdp {
+    fn fmt(&self, formatter: &mut fmt::Formatter<'_>) -> fmt::Result {
+        formatter.write_str("<redacted>")
+    }
+}
+
+impl Serialize for RealtimeOfferSdp {
+    fn serialize<S>(&self, serializer: S) -> Result<S::Ok, S::Error>
+    where
+        S: Serializer,
+    {
+        serializer.serialize_str("<redacted>")
+    }
+}
+
+/// Keeps spoken answers out of command diagnostics and session recordings.
+/// Convert back to a string only at the app-server signaling boundary.
+#[derive(Clone, PartialEq, Eq)]
+pub(crate) struct RealtimeSpeechText(String);
+
+impl From<String> for RealtimeSpeechText {
+    fn from(value: String) -> Self {
+        Self(value)
+    }
+}
+
+impl From<RealtimeSpeechText> for String {
+    fn from(value: RealtimeSpeechText) -> Self {
+        value.0
+    }
+}
+
+impl RealtimeSpeechText {
+    pub(crate) fn as_str(&self) -> &str {
+        &self.0
+    }
+}
+
+impl fmt::Debug for RealtimeSpeechText {
+    fn fmt(&self, formatter: &mut fmt::Formatter<'_>) -> fmt::Result {
+        formatter.write_str("<redacted>")
+    }
+}
+
+impl Serialize for RealtimeSpeechText {
+    fn serialize<S>(&self, serializer: S) -> Result<S::Ok, S::Error>
+    where
+        S: Serializer,
+    {
+        serializer.serialize_str("<redacted>")
+    }
+}
+
 #[allow(clippy::large_enum_variant)]
 #[derive(Debug, Clone, PartialEq, Serialize)]
 pub(crate) enum AppCommand {
     Interrupt,
     CleanBackgroundTerminals,
+    RealtimeConversationStart {
+        thread_id: ThreadId,
+        offer_sdp: RealtimeOfferSdp,
+    },
+    RealtimeConversationStop {
+        thread_id: ThreadId,
+    },
+    RealtimeConversationSpeech {
+        thread_id: ThreadId,
+        attempt_id: u64,
+        input_generation: u64,
+        delivery_id: u64,
+        text: RealtimeSpeechText,
+    },
     RunUserShellCommand {
         command: String,
     },
@@ -252,3 +339,7 @@ impl From<&AppCommand> for AppCommand {
         value.clone()
     }
 }
+
+#[cfg(test)]
+#[path = "app_command_tests.rs"]
+mod tests;