@telnyx/ai-agent-lib 0.2.2 → 0.2.4

package/dist/react/client-context.d.ts

@@ -2,4 +2,4 @@ import type { PropsWithChildren } from "react";
 import type { TelnyxAIAgentConstructorParams } from "../client";
 import { TelnyxAIAgent } from "../client";
 export declare const ClientContext: import("react").Context<TelnyxAIAgent | null>;
-export declare const TelnyxAIAgentProvider: ({ children, agentId, environment, versionId, debug, }: PropsWithChildren & TelnyxAIAgentConstructorParams) => import("react/jsx-runtime").JSX.Element;
+export declare const TelnyxAIAgentProvider: ({ children, agentId, environment, versionId, debug, vad, }: PropsWithChildren & TelnyxAIAgentConstructorParams) => import("react/jsx-runtime").JSX.Element;

package/dist/types.d.ts

@@ -13,8 +13,10 @@ export type TranscriptItem = {
 };
 export type AgentStateData = {
     state: AgentState;
-    /** Round-trip latency in milliseconds from when user stopped speaking until agent response began. Only present when state is "speaking". */
-    latencyMs?: number;
+    /** Latency in milliseconds from when user stopped speaking until agent response began. Only present when state is "speaking". */
+    userPerceivedLatencyMs?: number;
+    /** Latency in milliseconds for the initial agent greeting (first speech). Only present on first "speaking" state. */
+    greetingLatencyMs?: number;
     /** UTC timestamp (ISO 8601) when user stopped speaking and thinking state began. Only present when state is "thinking". */
     thinkingStartedAt?: string;
 };
@@ -30,5 +32,36 @@ export type AIAgentEvents = {
 export type AgentState = "speaking" | "listening" | "thinking";
 export type TranscriptEvents = Pick<AIAgentEvents, "transcript.item">;
 export type AgentStateEvents = Pick<AIAgentEvents, "conversation.agent.state">;
+/**
+ * Configuration options for Voice Activity Detection (VAD).
+ * These thresholds control how the library detects user speech and measures latency.
+ */
+export type VADOptions = {
+    /**
+     * Volume threshold (0-255) for detecting speech.
+     * Audio levels above this value are considered speech.
+     * @default 10
+     */
+    volumeThreshold?: number;
+    /**
+     * Duration of silence (ms) before triggering "thinking" state.
+     * Lower values = faster response but may cut off natural pauses.
+     * Higher values = more tolerant of pauses but slower state transitions.
+     * @default 1000
+     */
+    silenceDurationMs?: number;
+    /**
+     * Minimum speech duration (ms) to count as real user speech.
+     * Filters out brief noise spikes. Short words like "yes"/"no" are ~150ms.
+     * @default 100
+     */
+    minSpeechDurationMs?: number;
+    /**
+     * Maximum latency (ms) to report. Values above this are considered
+     * stale (likely from old noise) and won't be reported.
+     * @default undefined (no limit)
+     */
+    maxLatencyMs?: number;
+};
 export type { INotification, Call, ICallOptions };
 export { SwEvent };

package/package.json

@@ -1,7 +1,7 @@
 {
   "name": "@telnyx/ai-agent-lib",
   "private": false,
-  "version": "0.2.2",
+  "version": "0.2.4",
   "type": "module",
   "main": "dist/index.js",
   "types": "dist/index.d.ts",
@@ -21,7 +21,7 @@
     "react-dom": "^19.1.0"
   },
   "dependencies": {
-    "@telnyx/webrtc": "^2.25.12",
+    "@telnyx/webrtc": "^2.25.14",
     "eventemitter3": "^5.0.1",
     "jotai": "^2.12.5",
     "loglevel": "^1.9.2"