@openclaw/voice-call 2026.8.1-beta.2 → 2026.8.1-beta.3

package/dist/runtime-entry.js

@@ -1,2 +1,2 @@
-import { t as createVoiceCallRuntime } from "./runtime-entry-F13eJ4pp.js";
+import { t as createVoiceCallRuntime } from "./runtime-entry-DK8kPCGC.js";
 export { createVoiceCallRuntime };

package/dist/{telnyx-sIS_WgnY.js → telnyx-DO8dFQcb.js}

@@ -1,4 +1,4 @@
-import { s as verifyTelnyxWebhook, t as guardedJsonApiRequest } from "./guarded-json-api-DEDwgf4J.js";
+import { s as verifyTelnyxWebhook, t as guardedJsonApiRequest } from "./guarded-json-api-j4OEG0fR.js";
 import crypto from "node:crypto";
 //#region extensions/voice-call/src/providers/telnyx.ts
 function normalizeTelnyxDirection(direction) {
@@ -122,13 +122,17 @@ var TelnyxProvider = class {
 				type: "call.speaking",
 				text: data.payload?.text || ""
 			};
-			case "call.transcription": return {
-				...baseEvent,
-				type: "call.speech",
-				transcript: data.payload?.transcription_data?.transcript ?? data.payload?.transcription ?? "",
-				isFinal: data.payload?.transcription_data?.is_final ?? data.payload?.is_final ?? true,
-				confidence: data.payload?.transcription_data?.confidence ?? data.payload?.confidence
-			};
+			case "call.transcription": {
+				const transcript = data.payload?.transcription_data?.transcript ?? data.payload?.transcription ?? "";
+				if (!transcript.trim()) return null;
+				return {
+					...baseEvent,
+					type: "call.speech",
+					transcript,
+					isFinal: data.payload?.transcription_data?.is_final ?? data.payload?.is_final ?? true,
+					confidence: data.payload?.transcription_data?.confidence ?? data.payload?.confidence
+				};
+			}
 			case "call.hangup": return {
 				...baseEvent,
 				type: "call.ended",

package/dist/{twilio-CbeVrVAl.js → twilio-JDt_1nlz.js}

@@ -1,13 +1,13 @@
 import { fetchWithSsrFGuard } from "./runtime-api.js";
 import "./api.js";
 import { n as requireSupportedTwilioApiHostname, r as resolveTwilioApiBaseUrl } from "./twilio-region-4fkgz3UG.js";
-import { _ as escapeXml, a as normalizeProviderStatus, i as mapProviderStatusToEndReason, p as chunkAudio, r as isProviderStatusTerminal, s as getHeader, v as mapVoiceToPolly } from "./runtime-entry-F13eJ4pp.js";
-import { c as verifyTwilioWebhook, i as readVoiceCallProviderJsonResponse, n as cancelProviderResponseBody, r as readProviderErrorResponseSnippet, t as guardedJsonApiRequest } from "./guarded-json-api-DEDwgf4J.js";
+import { _ as escapeXml, a as normalizeProviderStatus, i as mapProviderStatusToEndReason, p as chunkAudio, r as isProviderStatusTerminal, s as getHeader, v as mapVoiceToPolly } from "./runtime-entry-DK8kPCGC.js";
+import { c as verifyTwilioWebhook, i as readVoiceCallProviderJsonResponse, n as cancelProviderResponseBody, r as readProviderErrorResponseSnippet, t as guardedJsonApiRequest } from "./guarded-json-api-j4OEG0fR.js";
 import { normalizeOptionalString } from "openclaw/plugin-sdk/string-coerce-runtime";
 import { sleepWithAbort } from "openclaw/plugin-sdk/runtime-env";
 import crypto from "node:crypto";
 import { safeEqualSecret } from "openclaw/plugin-sdk/security-runtime";
-import { setTimeout as setTimeout$1 } from "node:timers/promises";
+import { retryAsync } from "openclaw/plugin-sdk/retry-runtime";
 //#region extensions/voice-call/src/providers/twilio/api.ts
 const TWILIO_API_TIMEOUT_MS = 3e4;
 /** Parse Twilio JSON error responses without trusting response shape. */
@@ -188,6 +188,7 @@ var TwilioProvider = class TwilioProvider {
 		this.currentPublicUrl = null;
 		this.ttsProvider = null;
 		this.mediaStreamHandler = null;
+		this.playbackMarkSequence = 0;
 		this.callStreamMap = /* @__PURE__ */ new Map();
 		this.streamAuthTokens = /* @__PURE__ */ new Map();
 		this.twimlStorage = /* @__PURE__ */ new Map();
@@ -261,15 +262,16 @@ var TwilioProvider = class TwilioProvider {
 		});
 	}
 	async updateLiveCallTwiml(providerCallId, twiml, operation) {
-		for (const retryDelayMs of TWILIO_CALL_UPDATE_RETRY_DELAYS_MS) try {
-			await this.apiRequest(`/Calls/${providerCallId}.json`, { Twiml: twiml });
-			return;
-		} catch (err) {
-			if (!isTwilioCallNotInProgressError(err)) throw err;
-			console.warn(`[voice-call] Twilio ${operation} update hit call state race (21220); retrying in ${retryDelayMs}ms`);
-			await setTimeout$1(retryDelayMs);
-		}
-		await this.apiRequest(`/Calls/${providerCallId}.json`, { Twiml: twiml });
+		await retryAsync(() => this.apiRequest(`/Calls/${providerCallId}.json`, { Twiml: twiml }), {
+			attempts: TWILIO_CALL_UPDATE_RETRY_DELAYS_MS.length + 1,
+			minDelayMs: 0,
+			shouldRetry: isTwilioCallNotInProgressError,
+			delayMs: ({ attempt }) => TWILIO_CALL_UPDATE_RETRY_DELAYS_MS[attempt - 1] ?? 0,
+			onRetry: ({ delayMs }) => {
+				console.warn(`[voice-call] Twilio ${operation} update hit call state race (21220); retrying in ${delayMs}ms`);
+			},
+			sleep: (delayMs) => sleepWithAbort(delayMs)
+		});
 	}
 	/**
 	* Verify Twilio webhook signature using HMAC-SHA1.
@@ -301,6 +303,7 @@ var TwilioProvider = class TwilioProvider {
 				dedupeKey,
 				turnToken: turnTokenFromQuery
 			});
+			if (event?.direction === "inbound" && event.type !== "call.ended" && event.providerCallId && this.currentPublicUrl && !this.callWebhookUrls.has(event.providerCallId)) this.callWebhookUrls.set(event.providerCallId, this.currentPublicUrl);
 			const twiml = this.generateTwimlResponse(ctx);
 			return {
 				events: event ? [event] : [],
@@ -345,7 +348,7 @@ var TwilioProvider = class TwilioProvider {
 			to: params.get("To") || void 0
 		};
 		const speechResult = params.get("SpeechResult");
-		if (speechResult) return {
+		if (speechResult?.trim()) return {
 			...baseEvent,
 			type: "call.speech",
 			transcript: speechResult,
@@ -587,10 +590,6 @@ var TwilioProvider = class TwilioProvider {
 			const raw = handler.sendAudio(streamSid, audio);
 			return normalizeSendResult(raw);
 		};
-		const sendPlaybackMark = (name) => {
-			const raw = handler.sendMark(streamSid, name);
-			return normalizeSendResult(raw);
-		};
 		await handler.queueTts(streamSid, async (signal) => {
 			const sendKeepAlive = () => {
 				sendAudioChunk(SILENCE_CHUNK);
@@ -601,6 +600,7 @@ var TwilioProvider = class TwilioProvider {
 			}, CHUNK_DELAY_MS);
 			let muLawAudio;
 			let synthTimeout = null;
+			let removeAbortListener = () => {};
 			const synthTimeoutMs = ttsProvider.synthesisTimeoutMs;
 			try {
 				const synthPromise = ttsProvider.synthesizeForTelephony(text);
@@ -609,10 +609,23 @@ var TwilioProvider = class TwilioProvider {
 						reject(/* @__PURE__ */ new Error(`Telephony TTS synthesis timed out after ${synthTimeoutMs}ms`));
 					}, synthTimeoutMs);
 				});
-				muLawAudio = await Promise.race([synthPromise, timeoutPromise]);
+				const abortPromise = new Promise((_, reject) => {
+					const onAbort = () => {
+						reject(signal.reason instanceof Error ? signal.reason : /* @__PURE__ */ new Error("Telephony TTS synthesis aborted"));
+					};
+					signal.addEventListener("abort", onAbort, { once: true });
+					removeAbortListener = () => signal.removeEventListener("abort", onAbort);
+					if (signal.aborted) onAbort();
+				});
+				muLawAudio = await Promise.race([
+					synthPromise,
+					timeoutPromise,
+					abortPromise
+				]);
 			} finally {
 				if (synthTimeout) clearTimeout(synthTimeout);
 				clearInterval(keepAlive);
+				removeAbortListener();
 			}
 			if (muLawAudio.length === 0) throw new Error("Telephony TTS produced no audio");
 			let chunkAttempts = 0;
@@ -621,7 +634,11 @@ var TwilioProvider = class TwilioProvider {
 			for (const chunk of chunkAudio(muLawAudio, CHUNK_SIZE)) {
 				if (signal.aborted) break;
 				chunkAttempts += 1;
-				if (sendAudioChunk(chunk).sent) chunkDelivered += 1;
+				if (!sendAudioChunk(chunk).sent) {
+					handler.clearAudio(streamSid);
+					throw new Error(`Telephony stream playback failed: audio chunk ${chunkAttempts} not delivered`);
+				}
+				chunkDelivered += 1;
 				const waitMs = nextChunkDueAt - Date.now();
 				if (waitMs > 0) try {
 					await sleepWithAbort(Math.ceil(waitMs), signal);
@@ -632,14 +649,10 @@ var TwilioProvider = class TwilioProvider {
 				nextChunkDueAt += CHUNK_DELAY_MS;
 				if (signal.aborted) break;
 			}
-			let markSent = true;
-			if (!signal.aborted) markSent = sendPlaybackMark(`tts-${Date.now()}`).sent;
-			if (!signal.aborted && chunkAttempts > 0 && (chunkDelivered === 0 || !markSent)) {
-				const failures = [];
-				if (chunkDelivered === 0) failures.push("no audio chunks delivered");
-				if (!markSent) failures.push("completion mark not delivered");
-				throw new Error(`Telephony stream playback failed: ${failures.join("; ")}`);
-			}
+			if (signal.aborted) return;
+			if (chunkAttempts === 0 || chunkDelivered !== chunkAttempts) throw new Error("Telephony stream playback failed: incomplete audio delivery");
+			const markName = `tts-${Date.now()}-${++this.playbackMarkSequence}`;
+			await handler.sendMarkAndWait(streamSid, markName, muLawAudio.length / 8, signal);
 		});
 	}
 	/**

package/openclaw.plugin.json

@@ -6,6 +6,13 @@
   },
   "name": "Voice Call",
   "description": "OpenClaw voice-call plugin for Twilio, Telnyx, and Plivo phone calls.",
+  "cliCommands": [
+    {
+      "name": "voicecall",
+      "description": "Voice call utilities",
+      "hasSubcommands": true
+    }
+  ],
   "skills": ["./skills"],
   "commandAliases": [{ "name": "voicecall" }],
   "activation": {
@@ -85,6 +92,11 @@
       "label": "Tailscale Mode",
       "advanced": true
     },
+    "tailscale.port": {
+      "label": "Tailscale HTTPS Port",
+      "help": "External HTTPS port. Funnel supports 443, 8443, or 10000; Serve accepts any valid port.",
+      "advanced": true
+    },
     "tailscale.path": {
       "label": "Tailscale Path",
       "advanced": true
@@ -235,7 +247,7 @@
     },
     "sessionScope": {
       "label": "Session Scope",
-      "help": "Use per-phone to preserve caller memory across calls, or per-call to isolate every call into a fresh voice session."
+      "help": "Use per-phone to preserve caller memory, per-call to isolate every call, or main to share the configured agent's main session."
     },
     "responseModel": {
       "label": "Response Model",
@@ -421,6 +433,12 @@
             "type": "string",
             "enum": ["off", "serve", "funnel"]
           },
+          "port": {
+            "type": "integer",
+            "minimum": 1,
+            "maximum": 65535,
+            "default": 443
+          },
           "path": {
             "type": "string"
           }
@@ -926,7 +944,7 @@
       },
       "sessionScope": {
         "type": "string",
-        "enum": ["per-phone", "per-call"]
+        "enum": ["per-phone", "per-call", "main"]
       },
       "responseModel": {
         "type": "string"

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@openclaw/voice-call",
-  "version": "2026.8.1-beta.2",
+  "version": "2026.8.1-beta.3",
   "description": "OpenClaw voice-call plugin for Twilio, Telnyx, and Plivo phone calls.",
   "repository": {
     "type": "git",
@@ -13,7 +13,7 @@
     "zod": "4.4.3"
   },
   "peerDependencies": {
-    "openclaw": ">=2026.8.1-beta.2"
+    "openclaw": ">=2026.8.1-beta.3"
   },
   "peerDependenciesMeta": {
     "openclaw": {
@@ -30,10 +30,10 @@
       "minHostVersion": ">=2026.4.10"
     },
     "compat": {
-      "pluginApi": ">=2026.8.1-beta.2"
+      "pluginApi": ">=2026.8.1-beta.3"
     },
     "build": {
-      "openclawVersion": "2026.8.1-beta.2"
+      "openclawVersion": "2026.8.1-beta.3"
     },
     "release": {
       "publishToClawHub": true,