@vellumai/plugin-api 0.12.2-staging.1 → 0.12.2-staging.3

package/index.d.ts

@@ -330,6 +330,12 @@ declare const AssistantConfigSchema: z.ZodObject<{
                 "your-own": "your-own";
             }>>;
         }, z.core.$strip>>;
+        "quickbooks-oauth": z.ZodDefault<z.ZodObject<{
+            mode: z.ZodDefault<z.ZodEnum<{
+                managed: "managed";
+                "your-own": "your-own";
+            }>>;
+        }, z.core.$strip>>;
         "eventbrite-oauth": z.ZodDefault<z.ZodObject<{
             mode: z.ZodDefault<z.ZodEnum<{
                 managed: "managed";
@@ -618,6 +624,7 @@ declare const AssistantConfigSchema: z.ZodObject<{
         fdPollIntervalMs: z.ZodDefault<z.ZodNumber>;
         highFdThresholdRatio: z.ZodDefault<z.ZodNumber>;
         fdWarnCooldownMs: z.ZodDefault<z.ZodNumber>;
+        mountWatchIntervalMs: z.ZodDefault<z.ZodNumber>;
         pluginSourceScanIntervalMs: z.ZodDefault<z.ZodNumber>;
         baselineSnapshotIntervalMs: z.ZodDefault<z.ZodNumber>;
     }, z.core.$strip>>;
@@ -5125,7 +5132,9 @@ export declare function listConversations(limit?: number, conversationType?: Con
  */
 export declare function listInstalledSkills(): Promise<ResolvedSkillEntry[]>;
 
-declare const _LIVE_VOICE_SERVER_FRAME_TYPES: readonly ["ready", "busy", "speech_started", "utterance_end", "utterance_discarded", "stt_partial", "stt_final", "thinking", "activity", "assistant_text_delta", "tts_audio", "tts_done", "turn_cancelled", "minimize_room", "metrics", "archived", "error"];
+declare const _LIVE_VOICE_SERVER_FRAME_TYPES: readonly ["ready", "busy", "speech_started", "utterance_end", "utterance_discarded", "stt_partial", "stt_final", "thinking", "activity", "assistant_text_delta", "tts_audio", "tts_done", "turn_cancelled", "minimize_room", "session_control", "metrics", "archived", "error"];
+
+declare const LIVE_VOICE_SESSION_CONTROLS: readonly ["end", "mute", "look_screen", "look_camera", "look_stop"];
 
 declare const LIVE_VOICE_TURN_DETECTION_MODES: readonly ["manual", "server_vad"];
 
@@ -5368,7 +5377,7 @@ declare interface LiveVoiceReadyServerFrame extends LiveVoiceServerFrameBase {
     readonly audioInput?: boolean;
 }
 
-export declare type LiveVoiceServerFrame = LiveVoiceReadyServerFrame | LiveVoiceBusyServerFrame | LiveVoiceSpeechStartedServerFrame | LiveVoiceUtteranceEndServerFrame | LiveVoiceUtteranceDiscardedServerFrame | LiveVoiceSttPartialServerFrame | LiveVoiceSttFinalServerFrame | LiveVoiceThinkingServerFrame | LiveVoiceActivityServerFrame | LiveVoiceAssistantTextDeltaServerFrame | LiveVoiceTtsAudioServerFrame | LiveVoiceTtsDoneServerFrame | LiveVoiceTurnCancelledServerFrame | LiveVoiceMinimizeRoomServerFrame | LiveVoiceMetricsServerFrame | LiveVoiceArchivedServerFrame | LiveVoiceErrorServerFrame;
+export declare type LiveVoiceServerFrame = LiveVoiceReadyServerFrame | LiveVoiceBusyServerFrame | LiveVoiceSpeechStartedServerFrame | LiveVoiceUtteranceEndServerFrame | LiveVoiceUtteranceDiscardedServerFrame | LiveVoiceSttPartialServerFrame | LiveVoiceSttFinalServerFrame | LiveVoiceThinkingServerFrame | LiveVoiceActivityServerFrame | LiveVoiceAssistantTextDeltaServerFrame | LiveVoiceTtsAudioServerFrame | LiveVoiceTtsDoneServerFrame | LiveVoiceTurnCancelledServerFrame | LiveVoiceMinimizeRoomServerFrame | LiveVoiceSessionControlServerFrame | LiveVoiceMetricsServerFrame | LiveVoiceArchivedServerFrame | LiveVoiceErrorServerFrame;
 
 declare interface LiveVoiceServerFrameBase {
     readonly type: LiveVoiceServerFrameType;
@@ -5379,6 +5388,32 @@ declare type LiveVoiceServerFrameType = (typeof _LIVE_VOICE_SERVER_FRAME_TYPES)[
 
 export declare type LiveVoiceSessionCloseReason = "client_end" | "error" | "websocket_close" | "transport_closed" | "client_timeout" | "forced_end" | "manager_shutdown";
 
+/** A session control a client can carry out on the assistant's behalf. */
+declare type LiveVoiceSessionControl = (typeof LIVE_VOICE_SESSION_CONTROLS)[number];
+
+/**
+ * A session control the just-completed reply asked for with a terminal marker
+ * (`[END_CALL]`, `[MUTE]`, `[MUTE:<seconds>]`). Sent only after the turn's TTS
+ * has fully drained, so the spoken acknowledgement is heard first, never for a
+ * turn the user barged in on, at most once per turn, and only for a control
+ * the client listed in the start frame's `sessionControls`.
+ *
+ * - `end`: end the session the way the client's own end control does.
+ * - `mute`: mute the microphone. With `durationMs`, unmute again once it
+ *   elapses; without, stay muted until the user unmutes. The timer is the
+ *   client's: a muted microphone sends silence, so the daemon cannot hear an
+ *   "unmute".
+ * - `look_screen`: start showing the call the user's screen.
+ * - `look_camera`: start showing the call what the camera sees.
+ * - `look_stop`: stop showing the call the screen and the camera.
+ */
+declare interface LiveVoiceSessionControlServerFrame extends LiveVoiceServerFrameBase {
+    readonly type: "session_control";
+    readonly turnId: string;
+    readonly action: LiveVoiceSessionControl;
+    readonly durationMs?: number;
+}
+
 /**
  * Where the session holding the daemon's single live-voice slot is running,
  * as much of it as the daemon knows.

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@vellumai/plugin-api",
-  "version": "0.12.2-staging.1",
+  "version": "0.12.2-staging.3",
   "description": "Public TypeScript authoring contract for Vellum assistant plugins.",
   "license": "MIT",
   "type": "module",