@xopcai/xopc 0.0.86 → 0.0.88

package/dist/src/voice/stt/list-providers.js

@@ -1,19 +1,54 @@
 import { listMediaUnderstandingProviders } from "../../media-understanding/registry.js";
 import { resolveSTTProviderConfig } from "./factory.js";
 import { mergeSttConfigFromAppConfig } from "../../channels/attachments/voice-stt-webchat.js";
+import { getVoiceProviderMetadata } from "../metadata/registry.js";
+import "../metadata/index.js";
 //#region src/voice/stt/list-providers.ts
+function fallbackMetadata(providerId, aliases) {
+	return {
+		id: providerId,
+		capability: "stt",
+		displayName: providerId,
+		aliases: [...aliases],
+		fields: [
+			{
+				key: "apiKey",
+				label: "API Key",
+				type: "password",
+				secret: true
+			},
+			{
+				key: "model",
+				label: "Model",
+				type: "string"
+			},
+			{
+				key: "baseUrl",
+				label: "Base URL",
+				type: "string"
+			}
+		],
+		diagnostics: {
+			requiresApiKey: true,
+			configPath: `tools.media.audio.providers.${providerId}`
+		}
+	};
+}
 function isSttProviderConfigured(providerId, config) {
 	return resolveSTTProviderConfig(providerId, config) !== null;
 }
-/** List registered audio STT providers with configured state for the current app config. */
+/** List registered audio STT providers with configured state and UI metadata. */
 function listSttProvidersForApi(config) {
 	const sttConfig = mergeSttConfigFromAppConfig(config?.tools?.media?.audio, config?.tools?.media);
 	return {
-		providers: listMediaUnderstandingProviders().filter((plugin) => plugin.capabilities?.includes("audio") && typeof plugin.transcribeAudio === "function").map((plugin) => ({
-			id: plugin.id,
-			aliases: [...plugin.aliases ?? []],
-			configured: isSttProviderConfigured(plugin.id, sttConfig)
-		})),
+		providers: listMediaUnderstandingProviders().filter((plugin) => plugin.capabilities?.includes("audio") && typeof plugin.transcribeAudio === "function").map((plugin) => {
+			const metadata = getVoiceProviderMetadata("stt", plugin.id) ?? fallbackMetadata(plugin.id, plugin.aliases ?? []);
+			return {
+				...metadata,
+				aliases: [...metadata.aliases ?? plugin.aliases ?? []],
+				configured: isSttProviderConfigured(plugin.id, sttConfig)
+			};
+		}),
 		active: sttConfig.provider
 	};
 }

package/dist/src/voice/stt/list-providers.js.map

@@ -1 +1 @@
-{"version":3,"file":"list-providers.js","names":[],"sources":["../../../../src/voice/stt/list-providers.ts"],"sourcesContent":["/**\n * Registry-driven STT provider listing for gateway / Web UI discovery.\n */\n\nimport type { Config } from '../../config/schema.js';\nimport { mergeSttConfigFromAppConfig } from '../../channels/attachments/voice-stt-webchat.js';\n\nimport { resolveSTTProviderConfig } from './factory.js';\nimport type { STTConfig } from './types.js';\nimport { listMediaUnderstandingProviders } from '../../media-understanding/registry.js';\n\nexport interface SttProviderListEntry {\n  id: string;\n  aliases: string[];\n  configured: boolean;\n}\n\nexport interface SttProvidersPayload {\n  providers: SttProviderListEntry[];\n  active: string;\n}\n\nexport function isSttProviderConfigured(providerId: string, config: STTConfig): boolean {\n  return resolveSTTProviderConfig(providerId, config) !== null;\n}\n\n/** List registered audio STT providers with configured state for the current app config. */\nexport function listSttProvidersForApi(config: Config | undefined): SttProvidersPayload {\n  const sttConfig = mergeSttConfigFromAppConfig(config?.tools?.media?.audio, config?.tools?.media);\n  const providers: SttProviderListEntry[] = listMediaUnderstandingProviders()\n    .filter(\n      (plugin) =>\n        plugin.capabilities?.includes('audio') && typeof plugin.transcribeAudio === 'function',\n    )\n    .map((plugin) => ({\n      id: plugin.id,\n      aliases: [...(plugin.aliases ?? [])],\n      configured: isSttProviderConfigured(plugin.id, sttConfig),\n    }));\n\n  return {\n    providers,\n    active: sttConfig.provider,\n  };\n}\n"],"mappings":";;;;AAsBA,SAAgB,wBAAwB,YAAoB,QAA4B;AACtF,QAAO,yBAAyB,YAAY,OAAO,KAAK;;;AAI1D,SAAgB,uBAAuB,QAAiD;CACtF,MAAM,YAAY,4BAA4B,QAAQ,OAAO,OAAO,OAAO,QAAQ,OAAO,MAAM;AAYhG,QAAO;EACL,WAZwC,iCAAiC,CACxE,QACE,WACC,OAAO,cAAc,SAAS,QAAQ,IAAI,OAAO,OAAO,oBAAoB,WAC/E,CACA,KAAK,YAAY;GAChB,IAAI,OAAO;GACX,SAAS,CAAC,GAAI,OAAO,WAAW,EAAE,CAAE;GACpC,YAAY,wBAAwB,OAAO,IAAI,UAAU;GAC1D,EAGQ;EACT,QAAQ,UAAU;EACnB"}
+{"version":3,"file":"list-providers.js","names":[],"sources":["../../../../src/voice/stt/list-providers.ts"],"sourcesContent":["/**\n * Registry-driven STT provider listing for gateway / Web UI discovery.\n */\n\nimport type { Config } from '../../config/schema.js';\nimport { mergeSttConfigFromAppConfig } from '../../channels/attachments/voice-stt-webchat.js';\nimport { listMediaUnderstandingProviders } from '../../media-understanding/registry.js';\nimport { getVoiceProviderMetadata, type VoiceProviderMetadata } from '../metadata/index.js';\n\nimport { resolveSTTProviderConfig } from './factory.js';\nimport type { STTConfig } from './types.js';\n\nexport interface SttProviderListEntry extends VoiceProviderMetadata {\n  aliases: string[];\n  configured: boolean;\n}\n\nexport interface SttProvidersPayload {\n  providers: SttProviderListEntry[];\n  active: string;\n}\n\nfunction fallbackMetadata(providerId: string, aliases: readonly string[]): VoiceProviderMetadata {\n  return {\n    id: providerId,\n    capability: 'stt',\n    displayName: providerId,\n    aliases: [...aliases],\n    fields: [\n      { key: 'apiKey', label: 'API Key', type: 'password', secret: true },\n      { key: 'model', label: 'Model', type: 'string' },\n      { key: 'baseUrl', label: 'Base URL', type: 'string' },\n    ],\n    diagnostics: {\n      requiresApiKey: true,\n      configPath: `tools.media.audio.providers.${providerId}`,\n    },\n  };\n}\n\nexport function isSttProviderConfigured(providerId: string, config: STTConfig): boolean {\n  return resolveSTTProviderConfig(providerId, config) !== null;\n}\n\n/** List registered audio STT providers with configured state and UI metadata. */\nexport function listSttProvidersForApi(config: Config | undefined): SttProvidersPayload {\n  const sttConfig = mergeSttConfigFromAppConfig(config?.tools?.media?.audio, config?.tools?.media);\n  const providers: SttProviderListEntry[] = listMediaUnderstandingProviders()\n    .filter(\n      (plugin) =>\n        plugin.capabilities?.includes('audio') && typeof plugin.transcribeAudio === 'function',\n    )\n    .map((plugin) => {\n      const metadata = getVoiceProviderMetadata('stt', plugin.id) ?? fallbackMetadata(plugin.id, plugin.aliases ?? []);\n      return {\n        ...metadata,\n        aliases: [...(metadata.aliases ?? plugin.aliases ?? [])],\n        configured: isSttProviderConfigured(plugin.id, sttConfig),\n      };\n    });\n\n  return {\n    providers,\n    active: sttConfig.provider,\n  };\n}\n"],"mappings":";;;;;;AAsBA,SAAS,iBAAiB,YAAoB,SAAmD;AAC/F,QAAO;EACL,IAAI;EACJ,YAAY;EACZ,aAAa;EACb,SAAS,CAAC,GAAG,QAAQ;EACrB,QAAQ;GACN;IAAE,KAAK;IAAU,OAAO;IAAW,MAAM;IAAY,QAAQ;IAAM;GACnE;IAAE,KAAK;IAAS,OAAO;IAAS,MAAM;IAAU;GAChD;IAAE,KAAK;IAAW,OAAO;IAAY,MAAM;IAAU;GACtD;EACD,aAAa;GACX,gBAAgB;GAChB,YAAY,+BAA+B;GAC5C;EACF;;AAGH,SAAgB,wBAAwB,YAAoB,QAA4B;AACtF,QAAO,yBAAyB,YAAY,OAAO,KAAK;;;AAI1D,SAAgB,uBAAuB,QAAiD;CACtF,MAAM,YAAY,4BAA4B,QAAQ,OAAO,OAAO,OAAO,QAAQ,OAAO,MAAM;AAehG,QAAO;EACL,WAfwC,iCAAiC,CACxE,QACE,WACC,OAAO,cAAc,SAAS,QAAQ,IAAI,OAAO,OAAO,oBAAoB,WAC/E,CACA,KAAK,WAAW;GACf,MAAM,WAAW,yBAAyB,OAAO,OAAO,GAAG,IAAI,iBAAiB,OAAO,IAAI,OAAO,WAAW,EAAE,CAAC;AAChH,UAAO;IACL,GAAG;IACH,SAAS,CAAC,GAAI,SAAS,WAAW,OAAO,WAAW,EAAE,CAAE;IACxD,YAAY,wBAAwB,OAAO,IAAI,UAAU;IAC1D;IAIM;EACT,QAAQ,UAAU;EACnB"}

package/dist/src/voice/stt/types.d.ts

@@ -27,27 +27,11 @@ import type { MediaUnderstandingModelEntry } from '../../media-understanding/typ
 export interface STTConfig {
     enabled: boolean;
     provider: string;
-    alibaba?: {
-        apiKey?: string;
-        model?: string;
-        baseUrl?: string;
-        headers?: Record<string, string>;
-        language?: string;
-        prompt?: string;
-    };
-    openai?: {
-        apiKey?: string;
-        model?: string;
-        baseUrl?: string;
-        headers?: Record<string, string>;
-        language?: string;
-        prompt?: string;
-    };
     /** Capability-local model chain (`tools.media.audio.models`). */
     models?: MediaUnderstandingModelEntry[];
     /** Shared model entries merged from `tools.media.models` at runtime. */
     sharedModels?: MediaUnderstandingModelEntry[];
-    /** OpenClaw-aligned provider settings map (`tools.media.audio.providers.<id>`). */
+    /** Provider settings map (`tools.media.audio.providers.<id>`). */
     providers?: Record<string, Record<string, unknown>>;
     fallback?: {
         enabled: boolean;
@@ -55,7 +39,6 @@ export interface STTConfig {
     };
     /** Hard timeout per provider call (ms). Default 60s. */
     timeoutMs?: number;
-    [key: string]: unknown;
 }
 export type STTProviderAttemptOutcome = 'success' | 'skipped' | 'failed';
 export type STTProviderFailureReason = 'success' | 'not_configured' | 'timeout' | 'provider_error' | 'unsupported_format' | 'unknown';

package/dist/src/voice/stt/types.js

@@ -2,8 +2,10 @@
 const DEFAULT_STT_CONFIG = {
 	enabled: false,
 	provider: "alibaba",
-	alibaba: { model: "paraformer-v2" },
-	openai: { model: "whisper-1" },
+	providers: {
+		alibaba: { model: "paraformer-v2" },
+		openai: { model: "whisper-1" }
+	},
 	fallback: {
 		enabled: true,
 		order: ["alibaba", "openai"]

package/dist/src/voice/stt/types.js.map

@@ -1 +1 @@
-{"version":3,"file":"types.js","names":[],"sources":["../../../../src/voice/stt/types.ts"],"sourcesContent":["/**\n * STT public types — data shapes consumed by downstream code (channels,\n * schema validation, gateway payloads).\n *\n * Provider implementations live behind `MediaUnderstandingProvider.transcribeAudio`\n * (see src/media-understanding/types.ts).\n */\n\nexport interface STTResult {\n  /** Transcribed text. */\n  text: string;\n  /** Provider that performed the transcription (e.g. \"alibaba\", \"openai\"). */\n  provider: string;\n  /** Audio duration in seconds (if reported). */\n  duration?: number;\n  /** Detected language (if reported). */\n  language?: string;\n}\n\nexport interface STTOptions {\n  /** Language hint (e.g. 'zh', 'en'). */\n  language?: string;\n  /** Model id (provider-specific). */\n  model?: string;\n}\n\nexport type { MediaUnderstandingModelEntry } from '../../media-understanding/types.js';\nimport type { MediaUnderstandingModelEntry } from '../../media-understanding/types.js';\n\n/** STTConfig consumed by the schema and downstream wiring. */\nexport interface STTConfig {\n  enabled: boolean;\n  provider: string;\n  alibaba?: {\n    apiKey?: string;\n    model?: string;\n    baseUrl?: string;\n    headers?: Record<string, string>;\n    language?: string;\n    prompt?: string;\n  };\n  openai?: {\n    apiKey?: string;\n    model?: string;\n    baseUrl?: string;\n    headers?: Record<string, string>;\n    language?: string;\n    prompt?: string;\n  };\n  /** Capability-local model chain (`tools.media.audio.models`). */\n  models?: MediaUnderstandingModelEntry[];\n  /** Shared model entries merged from `tools.media.models` at runtime. */\n  sharedModels?: MediaUnderstandingModelEntry[];\n  /** OpenClaw-aligned provider settings map (`tools.media.audio.providers.<id>`). */\n  providers?: Record<string, Record<string, unknown>>;\n  fallback?: {\n    enabled: boolean;\n    order: string[];\n  };\n  /** Hard timeout per provider call (ms). Default 60s. */\n  timeoutMs?: number;\n  [key: string]: unknown;\n}\n\nexport type STTProviderAttemptOutcome = 'success' | 'skipped' | 'failed';\n\nexport type STTProviderFailureReason =\n  | 'success'\n  | 'not_configured'\n  | 'timeout'\n  | 'provider_error'\n  | 'unsupported_format'\n  | 'unknown';\n\nexport interface STTProviderAttempt {\n  provider: string;\n  outcome: STTProviderAttemptOutcome;\n  reasonCode: STTProviderFailureReason;\n  latencyMs: number;\n  error?: string;\n}\n\nexport interface STTResultWithTracking extends STTResult {\n  attempts: STTProviderAttempt[];\n  fallbackFrom?: string;\n  attemptedProviders: string[];\n}\n\nexport const DEFAULT_STT_CONFIG: STTConfig = {\n  enabled: false,\n  provider: 'alibaba',\n  alibaba: {\n    model: 'paraformer-v2',\n  },\n  openai: {\n    model: 'whisper-1',\n  },\n  fallback: {\n    enabled: true,\n    order: ['alibaba', 'openai'],\n  },\n  timeoutMs: 60_000,\n};\n"],"mappings":";AAwFA,MAAa,qBAAgC;CAC3C,SAAS;CACT,UAAU;CACV,SAAS,EACP,OAAO,iBACR;CACD,QAAQ,EACN,OAAO,aACR;CACD,UAAU;EACR,SAAS;EACT,OAAO,CAAC,WAAW,SAAS;EAC7B;CACD,WAAW;CACZ"}
+{"version":3,"file":"types.js","names":[],"sources":["../../../../src/voice/stt/types.ts"],"sourcesContent":["/**\n * STT public types — data shapes consumed by downstream code (channels,\n * schema validation, gateway payloads).\n *\n * Provider implementations live behind `MediaUnderstandingProvider.transcribeAudio`\n * (see src/media-understanding/types.ts).\n */\n\nexport interface STTResult {\n  /** Transcribed text. */\n  text: string;\n  /** Provider that performed the transcription (e.g. \"alibaba\", \"openai\"). */\n  provider: string;\n  /** Audio duration in seconds (if reported). */\n  duration?: number;\n  /** Detected language (if reported). */\n  language?: string;\n}\n\nexport interface STTOptions {\n  /** Language hint (e.g. 'zh', 'en'). */\n  language?: string;\n  /** Model id (provider-specific). */\n  model?: string;\n}\n\nexport type { MediaUnderstandingModelEntry } from '../../media-understanding/types.js';\nimport type { MediaUnderstandingModelEntry } from '../../media-understanding/types.js';\n\n/** STTConfig consumed by the schema and downstream wiring. */\nexport interface STTConfig {\n  enabled: boolean;\n  provider: string;\n  /** Capability-local model chain (`tools.media.audio.models`). */\n  models?: MediaUnderstandingModelEntry[];\n  /** Shared model entries merged from `tools.media.models` at runtime. */\n  sharedModels?: MediaUnderstandingModelEntry[];\n  /** Provider settings map (`tools.media.audio.providers.<id>`). */\n  providers?: Record<string, Record<string, unknown>>;\n  fallback?: {\n    enabled: boolean;\n    order: string[];\n  };\n  /** Hard timeout per provider call (ms). Default 60s. */\n  timeoutMs?: number;\n}\n\nexport type STTProviderAttemptOutcome = 'success' | 'skipped' | 'failed';\n\nexport type STTProviderFailureReason =\n  | 'success'\n  | 'not_configured'\n  | 'timeout'\n  | 'provider_error'\n  | 'unsupported_format'\n  | 'unknown';\n\nexport interface STTProviderAttempt {\n  provider: string;\n  outcome: STTProviderAttemptOutcome;\n  reasonCode: STTProviderFailureReason;\n  latencyMs: number;\n  error?: string;\n}\n\nexport interface STTResultWithTracking extends STTResult {\n  attempts: STTProviderAttempt[];\n  fallbackFrom?: string;\n  attemptedProviders: string[];\n}\n\nexport const DEFAULT_STT_CONFIG: STTConfig = {\n  enabled: false,\n  provider: 'alibaba',\n  providers: {\n    alibaba: { model: 'paraformer-v2' },\n    openai: { model: 'whisper-1' },\n  },\n  fallback: {\n    enabled: true,\n    order: ['alibaba', 'openai'],\n  },\n  timeoutMs: 60_000,\n};\n"],"mappings":";AAuEA,MAAa,qBAAgC;CAC3C,SAAS;CACT,UAAU;CACV,WAAW;EACT,SAAS,EAAE,OAAO,iBAAiB;EACnC,QAAQ,EAAE,OAAO,aAAa;EAC/B;CACD,UAAU;EACR,SAAS;EACT,OAAO,CAAC,WAAW,SAAS;EAC7B;CACD,WAAW;CACZ"}

package/dist/src/voice/tts/audio.js

@@ -1,7 +1,7 @@
 import { createLogger } from "../../utils/logger/index.js";
 import { init_logger } from "../../utils/logger.js";
-import { join } from "path";
 import { unlink, writeFile } from "fs/promises";
+import { join } from "path";
 import { spawn } from "child_process";
 import { tmpdir } from "os";
 //#region src/voice/tts/audio.ts

package/dist/src/voice/tts/config-slice.d.ts

@@ -1,19 +1,15 @@
 /**
  * TTS config slice resolution — maps persisted config to per-provider raw config.
  *
- * Supports OpenClaw-aligned `messages.tts.providers.<id>` plus legacy flat keys
- * (`messages.tts.openai`, `messages.tts.tts-local-cli`, …).
+ * Reads `messages.tts.providers.<id>` only — there is no legacy flat-key form.
  */
 import type { TTSConfig } from './types.js';
-/** Top-level `messages.tts` keys that are not provider config buckets. */
-export declare const TTS_CONFIG_RESERVED_KEYS: Set<string>;
-/** Collect provider-id → raw config entries from providers map + legacy flat keys. */
+/** Collect provider-id → raw config entries from the `providers` map. */
 export declare function collectTtsProviderConfigEntries(config: Partial<TTSConfig> | Record<string, unknown> | undefined): Record<string, Record<string, unknown>>;
 /** Resolve the raw config slice for one provider id. */
 export declare function resolveTtsProviderConfigSlice(providerId: string, config: Partial<TTSConfig> | Record<string, unknown> | undefined): Record<string, unknown>;
 /**
  * Build the `rawConfig` object passed into `SpeechProviderPlugin.resolveConfig`.
- * Includes the full TTS block, normalized `providers` map, and a top-level slice
- * for the active provider (matches built-in `rawConfig[id] ?? rawConfig` reads).
+ * Includes the full TTS block plus a normalized `providers` map.
  */
 export declare function buildTtsResolveRawConfig(providerId: string, config: Partial<TTSConfig> | Record<string, unknown>): Record<string, unknown>;

package/dist/src/voice/tts/config-slice.js

@@ -1,42 +1,13 @@
 //#region src/voice/tts/config-slice.ts
-/** Top-level `messages.tts` keys that are not provider config buckets. */
-const TTS_CONFIG_RESERVED_KEYS = new Set([
-	"auto",
-	"enabled",
-	"maxTextLength",
-	"mode",
-	"modelOverrides",
-	"persona",
-	"personas",
-	"prefsPath",
-	"provider",
-	"providers",
-	"summaryModel",
-	"timeoutMs",
-	"trigger",
-	"fallback",
-	"summarization"
-]);
 function asRecord(value) {
 	return typeof value === "object" && value !== null && !Array.isArray(value) ? value : void 0;
 }
-function asProviderConfig(value) {
-	return asRecord(value) ?? {};
-}
-/** Collect provider-id → raw config entries from providers map + legacy flat keys. */
+/** Collect provider-id → raw config entries from the `providers` map. */
 function collectTtsProviderConfigEntries(config) {
-	const raw = config ?? {};
+	const providers = asRecord((config ?? {}).providers);
+	if (!providers) return {};
 	const entries = {};
-	const providers = asRecord(raw.providers);
-	if (providers) for (const [providerId, value] of Object.entries(providers)) entries[providerId] = {
-		...entries[providerId],
-		...asProviderConfig(value)
-	};
-	for (const [key, value] of Object.entries(raw)) {
-		if (TTS_CONFIG_RESERVED_KEYS.has(key)) continue;
-		if (typeof value !== "object" || value === null || Array.isArray(value)) continue;
-		if (entries[key] === void 0) entries[key] = asProviderConfig(value);
-	}
+	for (const [providerId, value] of Object.entries(providers)) entries[providerId] = { ...asRecord(value) ?? {} };
 	return entries;
 }
 /** Resolve the raw config slice for one provider id. */
@@ -45,19 +16,17 @@ function resolveTtsProviderConfigSlice(providerId, config) {
 }
 /**
 * Build the `rawConfig` object passed into `SpeechProviderPlugin.resolveConfig`.
-* Includes the full TTS block, normalized `providers` map, and a top-level slice
-* for the active provider (matches built-in `rawConfig[id] ?? rawConfig` reads).
+* Includes the full TTS block plus a normalized `providers` map.
 */
 function buildTtsResolveRawConfig(providerId, config) {
 	const entries = collectTtsProviderConfigEntries(config);
-	const slice = entries[providerId] ?? {};
 	return {
 		...config,
 		providers: entries,
-		[providerId]: slice
+		[providerId]: entries[providerId] ?? {}
 	};
 }
 //#endregion
-export { TTS_CONFIG_RESERVED_KEYS, buildTtsResolveRawConfig, collectTtsProviderConfigEntries, resolveTtsProviderConfigSlice };
+export { buildTtsResolveRawConfig, collectTtsProviderConfigEntries, resolveTtsProviderConfigSlice };
 
 //# sourceMappingURL=config-slice.js.map

package/dist/src/voice/tts/config-slice.js.map

@@ -1 +1 @@
-{"version":3,"file":"config-slice.js","names":[],"sources":["../../../../src/voice/tts/config-slice.ts"],"sourcesContent":["/**\n * TTS config slice resolution — maps persisted config to per-provider raw config.\n *\n * Supports OpenClaw-aligned `messages.tts.providers.<id>` plus legacy flat keys\n * (`messages.tts.openai`, `messages.tts.tts-local-cli`, …).\n */\n\nimport type { TTSConfig } from './types.js';\n\n/** Top-level `messages.tts` keys that are not provider config buckets. */\nexport const TTS_CONFIG_RESERVED_KEYS = new Set([\n  'auto',\n  'enabled',\n  'maxTextLength',\n  'mode',\n  'modelOverrides',\n  'persona',\n  'personas',\n  'prefsPath',\n  'provider',\n  'providers',\n  'summaryModel',\n  'timeoutMs',\n  'trigger',\n  'fallback',\n  'summarization',\n]);\n\nfunction asRecord(value: unknown): Record<string, unknown> | undefined {\n  return typeof value === 'object' && value !== null && !Array.isArray(value)\n    ? (value as Record<string, unknown>)\n    : undefined;\n}\n\nfunction asProviderConfig(value: unknown): Record<string, unknown> {\n  return asRecord(value) ?? {};\n}\n\n/** Collect provider-id → raw config entries from providers map + legacy flat keys. */\nexport function collectTtsProviderConfigEntries(\n  config: Partial<TTSConfig> | Record<string, unknown> | undefined,\n): Record<string, Record<string, unknown>> {\n  const raw = (config ?? {}) as Record<string, unknown>;\n  const entries: Record<string, Record<string, unknown>> = {};\n\n  const providers = asRecord(raw.providers);\n  if (providers) {\n    for (const [providerId, value] of Object.entries(providers)) {\n      entries[providerId] = { ...entries[providerId], ...asProviderConfig(value) };\n    }\n  }\n\n  for (const [key, value] of Object.entries(raw)) {\n    if (TTS_CONFIG_RESERVED_KEYS.has(key)) {\n      continue;\n    }\n    if (typeof value !== 'object' || value === null || Array.isArray(value)) {\n      continue;\n    }\n    if (entries[key] === undefined) {\n      entries[key] = asProviderConfig(value);\n    }\n  }\n\n  return entries;\n}\n\n/** Resolve the raw config slice for one provider id. */\nexport function resolveTtsProviderConfigSlice(\n  providerId: string,\n  config: Partial<TTSConfig> | Record<string, unknown> | undefined,\n): Record<string, unknown> {\n  const entries = collectTtsProviderConfigEntries(config);\n  return entries[providerId] ?? {};\n}\n\n/**\n * Build the `rawConfig` object passed into `SpeechProviderPlugin.resolveConfig`.\n * Includes the full TTS block, normalized `providers` map, and a top-level slice\n * for the active provider (matches built-in `rawConfig[id] ?? rawConfig` reads).\n */\nexport function buildTtsResolveRawConfig(\n  providerId: string,\n  config: Partial<TTSConfig> | Record<string, unknown>,\n): Record<string, unknown> {\n  const entries = collectTtsProviderConfigEntries(config);\n  const slice = entries[providerId] ?? {};\n  return {\n    ...(config as Record<string, unknown>),\n    providers: entries,\n    [providerId]: slice,\n  };\n}\n"],"mappings":";;AAUA,MAAa,2BAA2B,IAAI,IAAI;CAC9C;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACD,CAAC;AAEF,SAAS,SAAS,OAAqD;AACrE,QAAO,OAAO,UAAU,YAAY,UAAU,QAAQ,CAAC,MAAM,QAAQ,MAAM,GACtE,QACD,KAAA;;AAGN,SAAS,iBAAiB,OAAyC;AACjE,QAAO,SAAS,MAAM,IAAI,EAAE;;;AAI9B,SAAgB,gCACd,QACyC;CACzC,MAAM,MAAO,UAAU,EAAE;CACzB,MAAM,UAAmD,EAAE;CAE3D,MAAM,YAAY,SAAS,IAAI,UAAU;AACzC,KAAI,UACF,MAAK,MAAM,CAAC,YAAY,UAAU,OAAO,QAAQ,UAAU,CACzD,SAAQ,cAAc;EAAE,GAAG,QAAQ;EAAa,GAAG,iBAAiB,MAAM;EAAE;AAIhF,MAAK,MAAM,CAAC,KAAK,UAAU,OAAO,QAAQ,IAAI,EAAE;AAC9C,MAAI,yBAAyB,IAAI,IAAI,CACnC;AAEF,MAAI,OAAO,UAAU,YAAY,UAAU,QAAQ,MAAM,QAAQ,MAAM,CACrE;AAEF,MAAI,QAAQ,SAAS,KAAA,EACnB,SAAQ,OAAO,iBAAiB,MAAM;;AAI1C,QAAO;;;AAIT,SAAgB,8BACd,YACA,QACyB;AAEzB,QADgB,gCAAgC,OAClC,CAAC,eAAe,EAAE;;;;;;;AAQlC,SAAgB,yBACd,YACA,QACyB;CACzB,MAAM,UAAU,gCAAgC,OAAO;CACvD,MAAM,QAAQ,QAAQ,eAAe,EAAE;AACvC,QAAO;EACL,GAAI;EACJ,WAAW;GACV,aAAa;EACf"}
+{"version":3,"file":"config-slice.js","names":[],"sources":["../../../../src/voice/tts/config-slice.ts"],"sourcesContent":["/**\n * TTS config slice resolution — maps persisted config to per-provider raw config.\n *\n * Reads `messages.tts.providers.<id>` only — there is no legacy flat-key form.\n */\n\nimport type { TTSConfig } from './types.js';\n\nfunction asRecord(value: unknown): Record<string, unknown> | undefined {\n  return typeof value === 'object' && value !== null && !Array.isArray(value)\n    ? (value as Record<string, unknown>)\n    : undefined;\n}\n\n/** Collect provider-id → raw config entries from the `providers` map. */\nexport function collectTtsProviderConfigEntries(\n  config: Partial<TTSConfig> | Record<string, unknown> | undefined,\n): Record<string, Record<string, unknown>> {\n  const raw = (config ?? {}) as Record<string, unknown>;\n  const providers = asRecord(raw.providers);\n  if (!providers) return {};\n  const entries: Record<string, Record<string, unknown>> = {};\n  for (const [providerId, value] of Object.entries(providers)) {\n    entries[providerId] = { ...(asRecord(value) ?? {}) };\n  }\n  return entries;\n}\n\n/** Resolve the raw config slice for one provider id. */\nexport function resolveTtsProviderConfigSlice(\n  providerId: string,\n  config: Partial<TTSConfig> | Record<string, unknown> | undefined,\n): Record<string, unknown> {\n  return collectTtsProviderConfigEntries(config)[providerId] ?? {};\n}\n\n/**\n * Build the `rawConfig` object passed into `SpeechProviderPlugin.resolveConfig`.\n * Includes the full TTS block plus a normalized `providers` map.\n */\nexport function buildTtsResolveRawConfig(\n  providerId: string,\n  config: Partial<TTSConfig> | Record<string, unknown>,\n): Record<string, unknown> {\n  const entries = collectTtsProviderConfigEntries(config);\n  return {\n    ...(config as Record<string, unknown>),\n    providers: entries,\n    [providerId]: entries[providerId] ?? {},\n  };\n}\n"],"mappings":";AAQA,SAAS,SAAS,OAAqD;AACrE,QAAO,OAAO,UAAU,YAAY,UAAU,QAAQ,CAAC,MAAM,QAAQ,MAAM,GACtE,QACD,KAAA;;;AAIN,SAAgB,gCACd,QACyC;CAEzC,MAAM,YAAY,UADL,UAAU,EAAE,EACM,UAAU;AACzC,KAAI,CAAC,UAAW,QAAO,EAAE;CACzB,MAAM,UAAmD,EAAE;AAC3D,MAAK,MAAM,CAAC,YAAY,UAAU,OAAO,QAAQ,UAAU,CACzD,SAAQ,cAAc,EAAE,GAAI,SAAS,MAAM,IAAI,EAAE,EAAG;AAEtD,QAAO;;;AAIT,SAAgB,8BACd,YACA,QACyB;AACzB,QAAO,gCAAgC,OAAO,CAAC,eAAe,EAAE;;;;;;AAOlE,SAAgB,yBACd,YACA,QACyB;CACzB,MAAM,UAAU,gCAAgC,OAAO;AACvD,QAAO;EACL,GAAI;EACJ,WAAW;GACV,aAAa,QAAQ,eAAe,EAAE;EACxC"}

package/dist/src/voice/tts/list-providers.d.ts

@@ -2,8 +2,8 @@
  * Registry-driven TTS provider listing for gateway / Web UI discovery.
  */
 import type { Config } from '../../config/schema.js';
-export interface TtsProviderListEntry {
-    id: string;
+import { type VoiceProviderMetadata } from '../metadata/index.js';
+export interface TtsProviderListEntry extends VoiceProviderMetadata {
     aliases: string[];
     configured: boolean;
 }
@@ -11,5 +11,5 @@ export interface TtsProvidersPayload {
     providers: TtsProviderListEntry[];
     active: string;
 }
-/** List registered speech providers with configured state for the current app config. */
+/** List registered speech providers with configured state and UI metadata. */
 export declare function listTtsProvidersForApi(config: Config | undefined): TtsProvidersPayload;

package/dist/src/voice/tts/list-providers.js

@@ -1,16 +1,51 @@
 import { listSpeechProviders } from "./speech-registry.js";
 import { resolveSpeechProvider } from "./factory.js";
 import { mergeTtsConfigFromAppConfig } from "./merge-config.js";
+import { getVoiceProviderMetadata } from "../metadata/registry.js";
+import "../metadata/index.js";
 //#region src/voice/tts/list-providers.ts
-/** List registered speech providers with configured state for the current app config. */
+function fallbackMetadata(providerId, aliases) {
+	return {
+		id: providerId,
+		capability: "tts",
+		displayName: providerId,
+		aliases: [...aliases],
+		fields: [
+			{
+				key: "apiKey",
+				label: "API Key",
+				type: "password",
+				secret: true
+			},
+			{
+				key: "model",
+				label: "Model",
+				type: "string"
+			},
+			{
+				key: "voice",
+				label: "Voice",
+				type: "string"
+			}
+		],
+		diagnostics: {
+			requiresApiKey: true,
+			configPath: `messages.tts.providers.${providerId}`
+		}
+	};
+}
+/** List registered speech providers with configured state and UI metadata. */
 function listTtsProvidersForApi(config) {
 	const ttsConfig = mergeTtsConfigFromAppConfig(config?.messages?.tts);
 	return {
-		providers: listSpeechProviders().map((plugin) => ({
-			id: plugin.id,
-			aliases: [...plugin.aliases ?? []],
-			configured: resolveSpeechProvider(plugin.id, ttsConfig) !== null
-		})),
+		providers: listSpeechProviders().map((plugin) => {
+			const metadata = getVoiceProviderMetadata("tts", plugin.id) ?? fallbackMetadata(plugin.id, plugin.aliases ?? []);
+			return {
+				...metadata,
+				aliases: [...metadata.aliases ?? plugin.aliases ?? []],
+				configured: resolveSpeechProvider(plugin.id, ttsConfig) !== null
+			};
+		}),
 		active: ttsConfig.provider
 	};
 }

package/dist/src/voice/tts/list-providers.js.map

@@ -1 +1 @@
-{"version":3,"file":"list-providers.js","names":[],"sources":["../../../../src/voice/tts/list-providers.ts"],"sourcesContent":["/**\n * Registry-driven TTS provider listing for gateway / Web UI discovery.\n */\n\nimport type { Config } from '../../config/schema.js';\n\nimport { resolveSpeechProvider } from './factory.js';\nimport { mergeTtsConfigFromAppConfig } from './merge-config.js';\nimport { listSpeechProviders } from './speech-registry.js';\n\nexport interface TtsProviderListEntry {\n  id: string;\n  aliases: string[];\n  configured: boolean;\n}\n\nexport interface TtsProvidersPayload {\n  providers: TtsProviderListEntry[];\n  active: string;\n}\n\n/** List registered speech providers with configured state for the current app config. */\nexport function listTtsProvidersForApi(config: Config | undefined): TtsProvidersPayload {\n  const ttsConfig = mergeTtsConfigFromAppConfig(config?.messages?.tts);\n  const providers: TtsProviderListEntry[] = listSpeechProviders().map((plugin) => ({\n    id: plugin.id,\n    aliases: [...(plugin.aliases ?? [])],\n    configured: resolveSpeechProvider(plugin.id, ttsConfig) !== null,\n  }));\n\n  return {\n    providers,\n    active: ttsConfig.provider,\n  };\n}\n"],"mappings":";;;;;AAsBA,SAAgB,uBAAuB,QAAiD;CACtF,MAAM,YAAY,4BAA4B,QAAQ,UAAU,IAAI;AAOpE,QAAO;EACL,WAPwC,qBAAqB,CAAC,KAAK,YAAY;GAC/E,IAAI,OAAO;GACX,SAAS,CAAC,GAAI,OAAO,WAAW,EAAE,CAAE;GACpC,YAAY,sBAAsB,OAAO,IAAI,UAAU,KAAK;GAC7D,EAGU;EACT,QAAQ,UAAU;EACnB"}
+{"version":3,"file":"list-providers.js","names":[],"sources":["../../../../src/voice/tts/list-providers.ts"],"sourcesContent":["/**\n * Registry-driven TTS provider listing for gateway / Web UI discovery.\n */\n\nimport type { Config } from '../../config/schema.js';\nimport { getVoiceProviderMetadata, type VoiceProviderMetadata } from '../metadata/index.js';\n\nimport { resolveSpeechProvider } from './factory.js';\nimport { mergeTtsConfigFromAppConfig } from './merge-config.js';\nimport { listSpeechProviders } from './speech-registry.js';\n\nexport interface TtsProviderListEntry extends VoiceProviderMetadata {\n  aliases: string[];\n  configured: boolean;\n}\n\nexport interface TtsProvidersPayload {\n  providers: TtsProviderListEntry[];\n  active: string;\n}\n\nfunction fallbackMetadata(providerId: string, aliases: readonly string[]): VoiceProviderMetadata {\n  return {\n    id: providerId,\n    capability: 'tts',\n    displayName: providerId,\n    aliases: [...aliases],\n    fields: [\n      { key: 'apiKey', label: 'API Key', type: 'password', secret: true },\n      { key: 'model', label: 'Model', type: 'string' },\n      { key: 'voice', label: 'Voice', type: 'string' },\n    ],\n    diagnostics: {\n      requiresApiKey: true,\n      configPath: `messages.tts.providers.${providerId}`,\n    },\n  };\n}\n\n/** List registered speech providers with configured state and UI metadata. */\nexport function listTtsProvidersForApi(config: Config | undefined): TtsProvidersPayload {\n  const ttsConfig = mergeTtsConfigFromAppConfig(config?.messages?.tts);\n  const providers: TtsProviderListEntry[] = listSpeechProviders().map((plugin) => {\n    const metadata = getVoiceProviderMetadata('tts', plugin.id) ?? fallbackMetadata(plugin.id, plugin.aliases ?? []);\n    return {\n      ...metadata,\n      aliases: [...(metadata.aliases ?? plugin.aliases ?? [])],\n      configured: resolveSpeechProvider(plugin.id, ttsConfig) !== null,\n    };\n  });\n\n  return {\n    providers,\n    active: ttsConfig.provider,\n  };\n}\n"],"mappings":";;;;;;AAqBA,SAAS,iBAAiB,YAAoB,SAAmD;AAC/F,QAAO;EACL,IAAI;EACJ,YAAY;EACZ,aAAa;EACb,SAAS,CAAC,GAAG,QAAQ;EACrB,QAAQ;GACN;IAAE,KAAK;IAAU,OAAO;IAAW,MAAM;IAAY,QAAQ;IAAM;GACnE;IAAE,KAAK;IAAS,OAAO;IAAS,MAAM;IAAU;GAChD;IAAE,KAAK;IAAS,OAAO;IAAS,MAAM;IAAU;GACjD;EACD,aAAa;GACX,gBAAgB;GAChB,YAAY,0BAA0B;GACvC;EACF;;;AAIH,SAAgB,uBAAuB,QAAiD;CACtF,MAAM,YAAY,4BAA4B,QAAQ,UAAU,IAAI;AAUpE,QAAO;EACL,WAVwC,qBAAqB,CAAC,KAAK,WAAW;GAC9E,MAAM,WAAW,yBAAyB,OAAO,OAAO,GAAG,IAAI,iBAAiB,OAAO,IAAI,OAAO,WAAW,EAAE,CAAC;AAChH,UAAO;IACL,GAAG;IACH,SAAS,CAAC,GAAI,SAAS,WAAW,OAAO,WAAW,EAAE,CAAE;IACxD,YAAY,sBAAsB,OAAO,IAAI,UAAU,KAAK;IAC7D;IAIQ;EACT,QAAQ,UAAU;EACnB"}

package/dist/src/voice/tts/merge-config.js

@@ -1,4 +1,3 @@
-import { TTS_CONFIG_RESERVED_KEYS, collectTtsProviderConfigEntries } from "./config-slice.js";
 import { isTTSAvailable } from "./factory.js";
 import { DEFAULT_TTS_CONFIG } from "./types.js";
 //#region src/voice/tts/merge-config.ts
@@ -23,51 +22,9 @@ function mergeProviderEntries(base, patch) {
 	};
 	return merged;
 }
-function mergeKnownFlatProviderSlices(merged, patch) {
-	return {
-		...merged,
-		alibaba: {
-			...DEFAULT_TTS_CONFIG.alibaba,
-			...patch.alibaba
-		},
-		openai: {
-			...DEFAULT_TTS_CONFIG.openai,
-			...patch.openai
-		},
-		edge: {
-			...DEFAULT_TTS_CONFIG.edge,
-			...patch.edge
-		},
-		minimax: {
-			...DEFAULT_TTS_CONFIG.minimax,
-			...patch.minimax
-		}
-	};
-}
-function mergeExtensionFlatProviderSlices(merged, patch) {
-	const next = { ...merged };
-	for (const [key, value] of Object.entries(patch)) {
-		if (TTS_CONFIG_RESERVED_KEYS.has(key)) continue;
-		if ([
-			"alibaba",
-			"openai",
-			"edge",
-			"minimax"
-		].includes(key)) continue;
-		if (typeof value !== "object" || value === null || Array.isArray(value)) continue;
-		const existing = next[key];
-		next[key] = typeof existing === "object" && existing !== null && !Array.isArray(existing) ? {
-			...existing,
-			...value
-		} : value;
-	}
-	return next;
-}
 function mergeTtsConfigFromAppConfig(tts) {
 	const p = tts ?? {};
-	const defaultEntries = collectTtsProviderConfigEntries(DEFAULT_TTS_CONFIG);
-	const patchEntries = collectTtsProviderConfigEntries(p);
-	let merged = {
+	return {
 		...DEFAULT_TTS_CONFIG,
 		...p,
 		enabled: p.enabled ?? DEFAULT_TTS_CONFIG.enabled,
@@ -81,15 +38,12 @@ function mergeTtsConfigFromAppConfig(tts) {
 			...DEFAULT_TTS_CONFIG.modelOverrides,
 			...p.modelOverrides
 		},
-		providers: mergeProviderEntries(defaultEntries, patchEntries),
+		providers: mergeProviderEntries(DEFAULT_TTS_CONFIG.providers, p.providers),
 		summarization: {
 			...DEFAULT_TTS_CONFIG.summarization,
 			...p.summarization
 		}
 	};
-	merged = mergeKnownFlatProviderSlices(merged, p);
-	merged = mergeExtensionFlatProviderSlices(merged, p);
-	return merged;
 }
 /**
 * User-facing hint when TTS is enabled in settings but no provider can run.

package/dist/src/voice/tts/merge-config.js.map

@@ -1 +1 @@
-{"version":3,"file":"merge-config.js","names":[],"sources":["../../../../src/voice/tts/merge-config.ts"],"sourcesContent":["import type { Config } from '../../config/schema.js';\nimport {\n  collectTtsProviderConfigEntries,\n  TTS_CONFIG_RESERVED_KEYS,\n} from './config-slice.js';\nimport { DEFAULT_TTS_CONFIG, type TTSConfig } from './types.js';\nimport { isTTSAvailable } from './factory.js';\n\n/**\n * Merge persisted app config `tts` with defaults to a full {@link TTSConfig}\n * for validation (provider chain, env-based keys, etc.).\n */\nfunction normalizeTtsTrigger(raw: unknown): TTSConfig['trigger'] {\n  const t = typeof raw === 'string' ? raw.toLowerCase() : '';\n  if (t === 'off' || t === 'always' || t === 'inbound' || t === 'tagged') return t;\n  return DEFAULT_TTS_CONFIG.trigger;\n}\n\nfunction normalizeTtsProvider(raw: unknown): string {\n  return typeof raw === 'string' && raw.trim().length > 0 ? raw.trim() : DEFAULT_TTS_CONFIG.provider;\n}\n\nfunction mergeProviderEntries(\n  base: Record<string, Record<string, unknown>> | undefined,\n  patch: Record<string, Record<string, unknown>> | undefined,\n): Record<string, Record<string, unknown>> | undefined {\n  if (!base && !patch) {\n    return undefined;\n  }\n  const merged: Record<string, Record<string, unknown>> = { ...(base ?? {}) };\n  for (const [providerId, slice] of Object.entries(patch ?? {})) {\n    merged[providerId] = { ...(merged[providerId] ?? {}), ...slice };\n  }\n  return merged;\n}\n\nfunction mergeKnownFlatProviderSlices(\n  merged: TTSConfig,\n  patch: Partial<TTSConfig>,\n): TTSConfig {\n  return {\n    ...merged,\n    alibaba: { ...DEFAULT_TTS_CONFIG.alibaba, ...patch.alibaba },\n    openai: { ...DEFAULT_TTS_CONFIG.openai, ...patch.openai },\n    edge: { ...DEFAULT_TTS_CONFIG.edge, ...patch.edge },\n    minimax: { ...DEFAULT_TTS_CONFIG.minimax, ...patch.minimax },\n  };\n}\n\nfunction mergeExtensionFlatProviderSlices(\n  merged: TTSConfig,\n  patch: Record<string, unknown>,\n): TTSConfig {\n  const next = { ...merged } as Record<string, unknown>;\n  for (const [key, value] of Object.entries(patch)) {\n    if (TTS_CONFIG_RESERVED_KEYS.has(key)) {\n      continue;\n    }\n    if (['alibaba', 'openai', 'edge', 'minimax'].includes(key)) {\n      continue;\n    }\n    if (typeof value !== 'object' || value === null || Array.isArray(value)) {\n      continue;\n    }\n    const existing = next[key];\n    next[key] =\n      typeof existing === 'object' && existing !== null && !Array.isArray(existing)\n        ? { ...(existing as Record<string, unknown>), ...(value as Record<string, unknown>) }\n        : value;\n  }\n  return next as TTSConfig;\n}\n\nexport function mergeTtsConfigFromAppConfig(tts: Partial<TTSConfig> | undefined): TTSConfig {\n  const p = (tts ?? {}) as Partial<TTSConfig> & Record<string, unknown>;\n  const defaultEntries = collectTtsProviderConfigEntries(DEFAULT_TTS_CONFIG);\n  const patchEntries = collectTtsProviderConfigEntries(p);\n\n  let merged: TTSConfig = {\n    ...DEFAULT_TTS_CONFIG,\n    ...p,\n    enabled: p.enabled ?? DEFAULT_TTS_CONFIG.enabled,\n    provider: normalizeTtsProvider(p.provider),\n    trigger: normalizeTtsTrigger(p.trigger ?? DEFAULT_TTS_CONFIG.trigger),\n    fallback: {\n      ...DEFAULT_TTS_CONFIG.fallback!,\n      ...p.fallback,\n    },\n    modelOverrides: {\n      ...DEFAULT_TTS_CONFIG.modelOverrides!,\n      ...p.modelOverrides,\n    },\n    providers: mergeProviderEntries(defaultEntries, patchEntries),\n    summarization: {\n      ...DEFAULT_TTS_CONFIG.summarization,\n      ...p.summarization,\n    },\n  };\n\n  merged = mergeKnownFlatProviderSlices(merged, p);\n  merged = mergeExtensionFlatProviderSlices(merged, p);\n  return merged;\n}\n\n/**\n * User-facing hint when TTS is enabled in settings but no provider can run.\n */\nexport function formatTtsSetupHint(): string {\n  return (\n    `⚠️ *TTS is on, but no provider can run yet.*\\n\\n` +\n    `Configure one of the following in \\`~/.xopc/xopc.json\\` (or env):\\n` +\n    `• *OpenAI*: \\`OPENAI_API_KEY\\` or \\`messages.tts.providers.openai.apiKey\\`\\n` +\n    `• *Alibaba*: \\`DASHSCOPE_API_KEY\\` or \\`messages.tts.providers.alibaba.apiKey\\`\\n` +\n    `• *MiniMax*: \\`MINIMAX_API_KEY\\` or \\`messages.tts.providers.minimax.apiKey\\`\\n` +\n    `• *Edge* (no key): ensure \\`messages.tts.providers.edge.enabled\\` is not \\`false\\`\\n` +\n    `• *Local CLI*: \\`messages.tts.providers.tts-local-cli.command\\`\\n\\n` +\n    `You can also use the gateway Web UI → Settings → Voice.`\n  );\n}\n\n/**\n * Append readiness / setup guidance when TTS is enabled but unavailable.\n */\nexport function appendTtsReadinessNote(content: string, appConfig: Config | undefined): string {\n  const effective = mergeTtsConfigFromAppConfig(appConfig?.messages?.tts);\n  if (!effective.enabled) {\n    return content;\n  }\n  if (isTTSAvailable(effective)) {\n    return content;\n  }\n  return `${content}\\n\\n${formatTtsSetupHint()}`;\n}\n"],"mappings":";;;;;;;;AAYA,SAAS,oBAAoB,KAAoC;CAC/D,MAAM,IAAI,OAAO,QAAQ,WAAW,IAAI,aAAa,GAAG;AACxD,KAAI,MAAM,SAAS,MAAM,YAAY,MAAM,aAAa,MAAM,SAAU,QAAO;AAC/E,QAAO,mBAAmB;;AAG5B,SAAS,qBAAqB,KAAsB;AAClD,QAAO,OAAO,QAAQ,YAAY,IAAI,MAAM,CAAC,SAAS,IAAI,IAAI,MAAM,GAAG,mBAAmB;;AAG5F,SAAS,qBACP,MACA,OACqD;AACrD,KAAI,CAAC,QAAQ,CAAC,MACZ;CAEF,MAAM,SAAkD,EAAE,GAAI,QAAQ,EAAE,EAAG;AAC3E,MAAK,MAAM,CAAC,YAAY,UAAU,OAAO,QAAQ,SAAS,EAAE,CAAC,CAC3D,QAAO,cAAc;EAAE,GAAI,OAAO,eAAe,EAAE;EAAG,GAAG;EAAO;AAElE,QAAO;;AAGT,SAAS,6BACP,QACA,OACW;AACX,QAAO;EACL,GAAG;EACH,SAAS;GAAE,GAAG,mBAAmB;GAAS,GAAG,MAAM;GAAS;EAC5D,QAAQ;GAAE,GAAG,mBAAmB;GAAQ,GAAG,MAAM;GAAQ;EACzD,MAAM;GAAE,GAAG,mBAAmB;GAAM,GAAG,MAAM;GAAM;EACnD,SAAS;GAAE,GAAG,mBAAmB;GAAS,GAAG,MAAM;GAAS;EAC7D;;AAGH,SAAS,iCACP,QACA,OACW;CACX,MAAM,OAAO,EAAE,GAAG,QAAQ;AAC1B,MAAK,MAAM,CAAC,KAAK,UAAU,OAAO,QAAQ,MAAM,EAAE;AAChD,MAAI,yBAAyB,IAAI,IAAI,CACnC;AAEF,MAAI;GAAC;GAAW;GAAU;GAAQ;GAAU,CAAC,SAAS,IAAI,CACxD;AAEF,MAAI,OAAO,UAAU,YAAY,UAAU,QAAQ,MAAM,QAAQ,MAAM,CACrE;EAEF,MAAM,WAAW,KAAK;AACtB,OAAK,OACH,OAAO,aAAa,YAAY,aAAa,QAAQ,CAAC,MAAM,QAAQ,SAAS,GACzE;GAAE,GAAI;GAAsC,GAAI;GAAmC,GACnF;;AAER,QAAO;;AAGT,SAAgB,4BAA4B,KAAgD;CAC1F,MAAM,IAAK,OAAO,EAAE;CACpB,MAAM,iBAAiB,gCAAgC,mBAAmB;CAC1E,MAAM,eAAe,gCAAgC,EAAE;CAEvD,IAAI,SAAoB;EACtB,GAAG;EACH,GAAG;EACH,SAAS,EAAE,WAAW,mBAAmB;EACzC,UAAU,qBAAqB,EAAE,SAAS;EAC1C,SAAS,oBAAoB,EAAE,WAAW,mBAAmB,QAAQ;EACrE,UAAU;GACR,GAAG,mBAAmB;GACtB,GAAG,EAAE;GACN;EACD,gBAAgB;GACd,GAAG,mBAAmB;GACtB,GAAG,EAAE;GACN;EACD,WAAW,qBAAqB,gBAAgB,aAAa;EAC7D,eAAe;GACb,GAAG,mBAAmB;GACtB,GAAG,EAAE;GACN;EACF;AAED,UAAS,6BAA6B,QAAQ,EAAE;AAChD,UAAS,iCAAiC,QAAQ,EAAE;AACpD,QAAO;;;;;AAMT,SAAgB,qBAA6B;AAC3C,QACE;;;;;AAcJ,SAAgB,uBAAuB,SAAiB,WAAuC;CAC7F,MAAM,YAAY,4BAA4B,WAAW,UAAU,IAAI;AACvE,KAAI,CAAC,UAAU,QACb,QAAO;AAET,KAAI,eAAe,UAAU,CAC3B,QAAO;AAET,QAAO,GAAG,QAAQ,MAAM,oBAAoB"}
+{"version":3,"file":"merge-config.js","names":[],"sources":["../../../../src/voice/tts/merge-config.ts"],"sourcesContent":["import type { Config } from '../../config/schema.js';\nimport { DEFAULT_TTS_CONFIG, type TTSConfig } from './types.js';\nimport { isTTSAvailable } from './factory.js';\n\n/**\n * Merge persisted app config `tts` with defaults to a full {@link TTSConfig}\n * for validation (provider chain, env-based keys, etc.).\n */\nfunction normalizeTtsTrigger(raw: unknown): TTSConfig['trigger'] {\n  const t = typeof raw === 'string' ? raw.toLowerCase() : '';\n  if (t === 'off' || t === 'always' || t === 'inbound' || t === 'tagged') return t;\n  return DEFAULT_TTS_CONFIG.trigger;\n}\n\nfunction normalizeTtsProvider(raw: unknown): string {\n  return typeof raw === 'string' && raw.trim().length > 0 ? raw.trim() : DEFAULT_TTS_CONFIG.provider;\n}\n\nfunction mergeProviderEntries(\n  base: Record<string, Record<string, unknown>> | undefined,\n  patch: Record<string, Record<string, unknown>> | undefined,\n): Record<string, Record<string, unknown>> | undefined {\n  if (!base && !patch) {\n    return undefined;\n  }\n  const merged: Record<string, Record<string, unknown>> = { ...(base ?? {}) };\n  for (const [providerId, slice] of Object.entries(patch ?? {})) {\n    merged[providerId] = { ...(merged[providerId] ?? {}), ...slice };\n  }\n  return merged;\n}\n\nexport function mergeTtsConfigFromAppConfig(tts: Partial<TTSConfig> | undefined): TTSConfig {\n  const p = (tts ?? {}) as Partial<TTSConfig>;\n  return {\n    ...DEFAULT_TTS_CONFIG,\n    ...p,\n    enabled: p.enabled ?? DEFAULT_TTS_CONFIG.enabled,\n    provider: normalizeTtsProvider(p.provider),\n    trigger: normalizeTtsTrigger(p.trigger ?? DEFAULT_TTS_CONFIG.trigger),\n    fallback: {\n      ...DEFAULT_TTS_CONFIG.fallback!,\n      ...p.fallback,\n    },\n    modelOverrides: {\n      ...DEFAULT_TTS_CONFIG.modelOverrides!,\n      ...p.modelOverrides,\n    },\n    providers: mergeProviderEntries(DEFAULT_TTS_CONFIG.providers, p.providers),\n    summarization: {\n      ...DEFAULT_TTS_CONFIG.summarization,\n      ...p.summarization,\n    },\n  };\n}\n\n/**\n * User-facing hint when TTS is enabled in settings but no provider can run.\n */\nexport function formatTtsSetupHint(): string {\n  return (\n    `⚠️ *TTS is on, but no provider can run yet.*\\n\\n` +\n    `Configure one of the following in \\`~/.xopc/xopc.json\\` (or env):\\n` +\n    `• *OpenAI*: \\`OPENAI_API_KEY\\` or \\`messages.tts.providers.openai.apiKey\\`\\n` +\n    `• *Alibaba*: \\`DASHSCOPE_API_KEY\\` or \\`messages.tts.providers.alibaba.apiKey\\`\\n` +\n    `• *MiniMax*: \\`MINIMAX_API_KEY\\` or \\`messages.tts.providers.minimax.apiKey\\`\\n` +\n    `• *Edge* (no key): ensure \\`messages.tts.providers.edge.enabled\\` is not \\`false\\`\\n` +\n    `• *Local CLI*: \\`messages.tts.providers.tts-local-cli.command\\`\\n\\n` +\n    `You can also use the gateway Web UI → Settings → Voice.`\n  );\n}\n\n/**\n * Append readiness / setup guidance when TTS is enabled but unavailable.\n */\nexport function appendTtsReadinessNote(content: string, appConfig: Config | undefined): string {\n  const effective = mergeTtsConfigFromAppConfig(appConfig?.messages?.tts);\n  if (!effective.enabled) {\n    return content;\n  }\n  if (isTTSAvailable(effective)) {\n    return content;\n  }\n  return `${content}\\n\\n${formatTtsSetupHint()}`;\n}\n"],"mappings":";;;;;;;AAQA,SAAS,oBAAoB,KAAoC;CAC/D,MAAM,IAAI,OAAO,QAAQ,WAAW,IAAI,aAAa,GAAG;AACxD,KAAI,MAAM,SAAS,MAAM,YAAY,MAAM,aAAa,MAAM,SAAU,QAAO;AAC/E,QAAO,mBAAmB;;AAG5B,SAAS,qBAAqB,KAAsB;AAClD,QAAO,OAAO,QAAQ,YAAY,IAAI,MAAM,CAAC,SAAS,IAAI,IAAI,MAAM,GAAG,mBAAmB;;AAG5F,SAAS,qBACP,MACA,OACqD;AACrD,KAAI,CAAC,QAAQ,CAAC,MACZ;CAEF,MAAM,SAAkD,EAAE,GAAI,QAAQ,EAAE,EAAG;AAC3E,MAAK,MAAM,CAAC,YAAY,UAAU,OAAO,QAAQ,SAAS,EAAE,CAAC,CAC3D,QAAO,cAAc;EAAE,GAAI,OAAO,eAAe,EAAE;EAAG,GAAG;EAAO;AAElE,QAAO;;AAGT,SAAgB,4BAA4B,KAAgD;CAC1F,MAAM,IAAK,OAAO,EAAE;AACpB,QAAO;EACL,GAAG;EACH,GAAG;EACH,SAAS,EAAE,WAAW,mBAAmB;EACzC,UAAU,qBAAqB,EAAE,SAAS;EAC1C,SAAS,oBAAoB,EAAE,WAAW,mBAAmB,QAAQ;EACrE,UAAU;GACR,GAAG,mBAAmB;GACtB,GAAG,EAAE;GACN;EACD,gBAAgB;GACd,GAAG,mBAAmB;GACtB,GAAG,EAAE;GACN;EACD,WAAW,qBAAqB,mBAAmB,WAAW,EAAE,UAAU;EAC1E,eAAe;GACb,GAAG,mBAAmB;GACtB,GAAG,EAAE;GACN;EACF;;;;;AAMH,SAAgB,qBAA6B;AAC3C,QACE;;;;;AAcJ,SAAgB,uBAAuB,SAAiB,WAAuC;CAC7F,MAAM,YAAY,4BAA4B,WAAW,UAAU,IAAI;AACvE,KAAI,CAAC,UAAU,QACb,QAAO;AAET,KAAI,eAAe,UAAU,CAC3B,QAAO;AAET,QAAO,GAAG,QAAQ,MAAM,oBAAoB"}

package/dist/src/voice/tts/providers/alibaba-speech.js

@@ -56,7 +56,7 @@ function trimToUndefined(value) {
 	return trimmed.length > 0 ? trimmed : void 0;
 }
 function normalizeConfig(rawConfig) {
-	const raw = asObject(rawConfig.alibaba) ?? rawConfig;
+	const raw = asObject(rawConfig.alibaba) ?? {};
 	return {
 		apiKey: trimToUndefined(raw.apiKey),
 		baseUrl: trimToUndefined(raw.baseUrl) ?? DEFAULT_BASE_URL,

package/dist/src/voice/tts/providers/alibaba-speech.js.map

@@ -1 +1 @@
-{"version":3,"file":"alibaba-speech.js","names":[],"sources":["../../../../../src/voice/tts/providers/alibaba-speech.ts"],"sourcesContent":["/**\n * Alibaba DashScope TTS provider (qwen-tts model).\n *\n * Implements the SpeechProviderPlugin contract directly because DashScope's\n * response shape (output.audio.url for the audio binary, no direct stream)\n * doesn't fit the OpenAI-compatible factory.\n *\n * Implementation notes (per docs/voice-rearchitecture.md §8.4.1):\n *   - `synthesizeStream` is intentionally not implemented. Native qwen-tts\n *     streaming uses a WebSocket protocol\n *     (wss://dashscope.aliyuncs.com/api-ws/v1/inference); speak-core's stream\n *     fallback wraps `synthesize` output as a single-chunk ReadableStream so\n *     callers see the same shape.\n *   - `maxTextLength = 512` (DashScope qwen-tts hard limit). Enforced upstream\n *     via `truncateAtSentenceBoundary` before this provider sees the text.\n *   - The audio URL returned by DashScope is hosted on Alibaba's CDN (variable\n *     hostname). We do not SSRF-guard the audio fetch because:\n *       (a) the URL is provided by the same provider we already trust for the\n *           initial synthesis call,\n *       (b) it's HTTPS-only and short-lived (~5 min TTL),\n *       (c) maintaining a hostname allowlist breaks every time Alibaba rotates\n *           CDN domains.\n *     A dedicated `assertSafeUrl` for trusted-CDN responses can be added later\n *     as a hardening pass.\n */\n\nimport {\n  ProviderHttpError,\n  fetchWithTimeoutGuarded,\n  postJsonRequest,\n} from '../../../media-shared/http/index.js';\nimport { createLogger } from '../../../utils/logger.js';\nimport { registerSpeechProvider } from '../speech-registry.js';\nimport type {\n  SpeechDirectiveTokenParseContext,\n  SpeechDirectiveTokenParseResult,\n  SpeechProviderConfig,\n  SpeechProviderConfiguredContext,\n  SpeechProviderPlugin,\n  SpeechProviderResolveConfigContext,\n  SpeechSynthesisRequest,\n  SpeechSynthesisResult,\n} from '../speech-provider-types.js';\n\nconst log = createLogger('SpeechProvider:Alibaba');\n\nconst DEFAULT_BASE_URL =\n  'https://dashscope.aliyuncs.com/api/v1/services/aigc/multimodal-generation/generation';\nconst DEFAULT_MODEL = 'qwen-tts';\nconst DEFAULT_VOICE = 'longxiaochun';\nconst ENV_KEY = 'DASHSCOPE_API_KEY';\nconst MAX_TEXT_LENGTH = 512;\nconst ALIBABA_VOICES = [\n  'Cherry',\n  'Ethan',\n  'Serena',\n  'Chelsie',\n  'longxiaochun',\n  'longxiaobai',\n  'longwan',\n  'longcheng',\n] as const;\n\ninterface AlibabaTtsConfig extends Record<string, unknown> {\n  apiKey?: string;\n  baseUrl: string;\n  model: string;\n  voice: string;\n}\n\ninterface CosyVoiceResponse {\n  output: {\n    audio?: { url?: string; data?: string };\n    speech?: string;\n    speech_url?: string;\n    finish_reason?: string;\n  };\n  usage?: { characters?: number };\n  request_id?: string;\n}\n\nfunction asObject(value: unknown): Record<string, unknown> | undefined {\n  return typeof value === 'object' && value !== null && !Array.isArray(value)\n    ? (value as Record<string, unknown>)\n    : undefined;\n}\n\nfunction trimToUndefined(value: unknown): string | undefined {\n  if (typeof value !== 'string') {\n    return undefined;\n  }\n  const trimmed = value.trim();\n  return trimmed.length > 0 ? trimmed : undefined;\n}\n\nfunction normalizeConfig(rawConfig: Record<string, unknown>): AlibabaTtsConfig {\n  const raw = asObject(rawConfig.alibaba) ?? rawConfig;\n  return {\n    apiKey: trimToUndefined(raw.apiKey),\n    baseUrl: trimToUndefined(raw.baseUrl) ?? DEFAULT_BASE_URL,\n    model: trimToUndefined(raw.model ?? raw.modelId) ?? DEFAULT_MODEL,\n    voice: trimToUndefined(raw.voice ?? raw.voiceId) ?? DEFAULT_VOICE,\n  };\n}\n\nfunction readProviderConfig(config: SpeechProviderConfig): AlibabaTtsConfig {\n  return {\n    apiKey: trimToUndefined(config.apiKey),\n    baseUrl: trimToUndefined(config.baseUrl) ?? DEFAULT_BASE_URL,\n    model: trimToUndefined(config.model ?? config.modelId) ?? DEFAULT_MODEL,\n    voice: trimToUndefined(config.voice ?? config.voiceId) ?? DEFAULT_VOICE,\n  };\n}\n\nfunction resolveApiKey(config: AlibabaTtsConfig): string | undefined {\n  return config.apiKey ?? trimToUndefined(process.env[ENV_KEY]);\n}\n\nfunction parseDirectiveTokenInternal(\n  ctx: SpeechDirectiveTokenParseContext,\n): SpeechDirectiveTokenParseResult {\n  switch (ctx.key) {\n    case 'voice':\n    case 'voice_id':\n    case 'voiceid':\n    case 'alibaba_voice':\n    case 'alibabavoice':\n      if (!ctx.policy.allowVoice) {\n        return { handled: true };\n      }\n      return { handled: true, overrides: { voice: ctx.value } };\n    case 'model':\n    case 'model_id':\n    case 'modelid':\n    case 'alibaba_model':\n    case 'alibabamodel':\n      if (!ctx.policy.allowModelId) {\n        return { handled: true };\n      }\n      return { handled: true, overrides: { model: ctx.value } };\n    default:\n      return { handled: false };\n  }\n}\n\nexport const alibabaSpeechProvider: SpeechProviderPlugin = {\n  id: 'alibaba',\n  aliases: ['dashscope', 'qwen-tts'],\n  autoSelectOrder: 25,\n\n  resolveConfig: (ctx: SpeechProviderResolveConfigContext) => normalizeConfig(ctx.rawConfig),\n\n  isConfigured: (ctx: SpeechProviderConfiguredContext) =>\n    Boolean(resolveApiKey(readProviderConfig(ctx.providerConfig))),\n\n  parseDirectiveToken: parseDirectiveTokenInternal,\n\n  listVoices: async () => ALIBABA_VOICES.map((id) => ({ id, name: id })),\n\n  synthesize: async (req: SpeechSynthesisRequest): Promise<SpeechSynthesisResult> => {\n    const config = readProviderConfig(req.providerConfig);\n    const apiKey = resolveApiKey(config);\n    if (!apiKey) {\n      throw new Error(\n        `Alibaba TTS API key missing (set ${ENV_KEY} or messages.tts.providers.alibaba.apiKey)`,\n      );\n    }\n    if (req.text.length > MAX_TEXT_LENGTH) {\n      throw new Error(\n        `Alibaba TTS text exceeds ${MAX_TEXT_LENGTH} char limit (got ${req.text.length})`,\n      );\n    }\n\n    const overrides = req.providerOverrides ?? {};\n    const model = trimToUndefined(overrides.model ?? overrides.modelId) ?? config.model;\n    const voice = trimToUndefined(overrides.voice ?? overrides.voiceId) ?? config.voice;\n\n    log.debug({ model, voice, textLength: req.text.length }, 'Calling Alibaba TTS');\n\n    const response = await postJsonRequest(config.baseUrl, {\n      timeoutMs: req.timeoutMs,\n      label: 'Alibaba TTS',\n      headers: {\n        Authorization: `Bearer ${apiKey}`,\n        'X-DashScope-DataInspection': 'disable',\n      },\n      body: {\n        model,\n        input: { text: req.text },\n        parameters: { voice },\n      },\n    });\n\n    const data = (await response.json()) as CosyVoiceResponse;\n\n    // DashScope occasionally returns `finish_reason: \"null\"` (string) on failure.\n    if (data.output?.finish_reason === 'null' && !data.output?.audio?.url) {\n      throw new Error(`Alibaba TTS API error: ${JSON.stringify(data)}`);\n    }\n\n    let audioBuffer: Buffer;\n    const audioUrl = data.output?.audio?.url ?? data.output?.speech_url;\n    if (audioUrl) {\n      // CDN URL — see file-level DECISION on why we skip SSRF here.\n      const audioResponse = await fetchWithTimeoutGuarded(audioUrl, {\n        timeoutMs: req.timeoutMs,\n        label: 'Alibaba TTS audio download',\n        allowPrivateNetwork: false,\n      });\n      if (!audioResponse.ok) {\n        throw new ProviderHttpError({\n          label: 'Alibaba TTS audio download',\n          status: audioResponse.status,\n          detail: audioResponse.statusText,\n        });\n      }\n      audioBuffer = Buffer.from(await audioResponse.arrayBuffer());\n    } else if (data.output?.speech) {\n      audioBuffer = Buffer.from(data.output.speech, 'base64');\n    } else if (data.output?.audio?.data) {\n      audioBuffer = Buffer.from(data.output.audio.data, 'base64');\n    } else {\n      throw new Error('No audio returned from Alibaba TTS');\n    }\n\n    log.debug(\n      { size: audioBuffer.length, characters: data.usage?.characters, requestId: data.request_id },\n      'Alibaba TTS completed',\n    );\n\n    return {\n      audioBuffer,\n      outputFormat: 'wav',\n      fileExtension: 'wav',\n      // wav is NOT a Telegram voice-note format → ffmpeg compresses downstream.\n      voiceCompatible: false,\n    };\n  },\n\n  // synthesizeStream intentionally omitted — see file-level DECISION.\n};\n\nregisterSpeechProvider(alibabaSpeechProvider);\n"],"mappings":";;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;aA+BwD;AAaxD,MAAM,MAAM,aAAa,yBAAyB;AAElD,MAAM,mBACJ;AACF,MAAM,gBAAgB;AACtB,MAAM,gBAAgB;AACtB,MAAM,UAAU;AAChB,MAAM,kBAAkB;AACxB,MAAM,iBAAiB;CACrB;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACD;AAoBD,SAAS,SAAS,OAAqD;AACrE,QAAO,OAAO,UAAU,YAAY,UAAU,QAAQ,CAAC,MAAM,QAAQ,MAAM,GACtE,QACD,KAAA;;AAGN,SAAS,gBAAgB,OAAoC;AAC3D,KAAI,OAAO,UAAU,SACnB;CAEF,MAAM,UAAU,MAAM,MAAM;AAC5B,QAAO,QAAQ,SAAS,IAAI,UAAU,KAAA;;AAGxC,SAAS,gBAAgB,WAAsD;CAC7E,MAAM,MAAM,SAAS,UAAU,QAAQ,IAAI;AAC3C,QAAO;EACL,QAAQ,gBAAgB,IAAI,OAAO;EACnC,SAAS,gBAAgB,IAAI,QAAQ,IAAI;EACzC,OAAO,gBAAgB,IAAI,SAAS,IAAI,QAAQ,IAAI;EACpD,OAAO,gBAAgB,IAAI,SAAS,IAAI,QAAQ,IAAI;EACrD;;AAGH,SAAS,mBAAmB,QAAgD;AAC1E,QAAO;EACL,QAAQ,gBAAgB,OAAO,OAAO;EACtC,SAAS,gBAAgB,OAAO,QAAQ,IAAI;EAC5C,OAAO,gBAAgB,OAAO,SAAS,OAAO,QAAQ,IAAI;EAC1D,OAAO,gBAAgB,OAAO,SAAS,OAAO,QAAQ,IAAI;EAC3D;;AAGH,SAAS,cAAc,QAA8C;AACnE,QAAO,OAAO,UAAU,gBAAgB,QAAQ,IAAI,SAAS;;AAG/D,SAAS,4BACP,KACiC;AACjC,SAAQ,IAAI,KAAZ;EACE,KAAK;EACL,KAAK;EACL,KAAK;EACL,KAAK;EACL,KAAK;AACH,OAAI,CAAC,IAAI,OAAO,WACd,QAAO,EAAE,SAAS,MAAM;AAE1B,UAAO;IAAE,SAAS;IAAM,WAAW,EAAE,OAAO,IAAI,OAAO;IAAE;EAC3D,KAAK;EACL,KAAK;EACL,KAAK;EACL,KAAK;EACL,KAAK;AACH,OAAI,CAAC,IAAI,OAAO,aACd,QAAO,EAAE,SAAS,MAAM;AAE1B,UAAO;IAAE,SAAS;IAAM,WAAW,EAAE,OAAO,IAAI,OAAO;IAAE;EAC3D,QACE,QAAO,EAAE,SAAS,OAAO;;;AAI/B,MAAa,wBAA8C;CACzD,IAAI;CACJ,SAAS,CAAC,aAAa,WAAW;CAClC,iBAAiB;CAEjB,gBAAgB,QAA4C,gBAAgB,IAAI,UAAU;CAE1F,eAAe,QACb,QAAQ,cAAc,mBAAmB,IAAI,eAAe,CAAC,CAAC;CAEhE,qBAAqB;CAErB,YAAY,YAAY,eAAe,KAAK,QAAQ;EAAE;EAAI,MAAM;EAAI,EAAE;CAEtE,YAAY,OAAO,QAAgE;EACjF,MAAM,SAAS,mBAAmB,IAAI,eAAe;EACrD,MAAM,SAAS,cAAc,OAAO;AACpC,MAAI,CAAC,OACH,OAAM,IAAI,MACR,oCAAoC,QAAQ,4CAC7C;AAEH,MAAI,IAAI,KAAK,SAAS,gBACpB,OAAM,IAAI,MACR,4BAA4B,gBAAgB,mBAAmB,IAAI,KAAK,OAAO,GAChF;EAGH,MAAM,YAAY,IAAI,qBAAqB,EAAE;EAC7C,MAAM,QAAQ,gBAAgB,UAAU,SAAS,UAAU,QAAQ,IAAI,OAAO;EAC9E,MAAM,QAAQ,gBAAgB,UAAU,SAAS,UAAU,QAAQ,IAAI,OAAO;AAE9E,MAAI,MAAM;GAAE;GAAO;GAAO,YAAY,IAAI,KAAK;GAAQ,EAAE,sBAAsB;EAgB/E,MAAM,OAAQ,OAAM,MAdG,gBAAgB,OAAO,SAAS;GACrD,WAAW,IAAI;GACf,OAAO;GACP,SAAS;IACP,eAAe,UAAU;IACzB,8BAA8B;IAC/B;GACD,MAAM;IACJ;IACA,OAAO,EAAE,MAAM,IAAI,MAAM;IACzB,YAAY,EAAE,OAAO;IACtB;GACF,CAAC,EAE2B,MAAM;AAGnC,MAAI,KAAK,QAAQ,kBAAkB,UAAU,CAAC,KAAK,QAAQ,OAAO,IAChE,OAAM,IAAI,MAAM,0BAA0B,KAAK,UAAU,KAAK,GAAG;EAGnE,IAAI;EACJ,MAAM,WAAW,KAAK,QAAQ,OAAO,OAAO,KAAK,QAAQ;AACzD,MAAI,UAAU;GAEZ,MAAM,gBAAgB,MAAM,wBAAwB,UAAU;IAC5D,WAAW,IAAI;IACf,OAAO;IACP,qBAAqB;IACtB,CAAC;AACF,OAAI,CAAC,cAAc,GACjB,OAAM,IAAI,kBAAkB;IAC1B,OAAO;IACP,QAAQ,cAAc;IACtB,QAAQ,cAAc;IACvB,CAAC;AAEJ,iBAAc,OAAO,KAAK,MAAM,cAAc,aAAa,CAAC;aACnD,KAAK,QAAQ,OACtB,eAAc,OAAO,KAAK,KAAK,OAAO,QAAQ,SAAS;WAC9C,KAAK,QAAQ,OAAO,KAC7B,eAAc,OAAO,KAAK,KAAK,OAAO,MAAM,MAAM,SAAS;MAE3D,OAAM,IAAI,MAAM,qCAAqC;AAGvD,MAAI,MACF;GAAE,MAAM,YAAY;GAAQ,YAAY,KAAK,OAAO;GAAY,WAAW,KAAK;GAAY,EAC5F,wBACD;AAED,SAAO;GACL;GACA,cAAc;GACd,eAAe;GAEf,iBAAiB;GAClB;;CAIJ;AAED,uBAAuB,sBAAsB"}
+{"version":3,"file":"alibaba-speech.js","names":[],"sources":["../../../../../src/voice/tts/providers/alibaba-speech.ts"],"sourcesContent":["/**\n * Alibaba DashScope TTS provider (qwen-tts model).\n *\n * Implements the SpeechProviderPlugin contract directly because DashScope's\n * response shape (output.audio.url for the audio binary, no direct stream)\n * doesn't fit the OpenAI-compatible factory.\n *\n * Implementation notes (per docs/voice-rearchitecture.md §8.4.1):\n *   - `synthesizeStream` is intentionally not implemented. Native qwen-tts\n *     streaming uses a WebSocket protocol\n *     (wss://dashscope.aliyuncs.com/api-ws/v1/inference); speak-core's stream\n *     fallback wraps `synthesize` output as a single-chunk ReadableStream so\n *     callers see the same shape.\n *   - `maxTextLength = 512` (DashScope qwen-tts hard limit). Enforced upstream\n *     via `truncateAtSentenceBoundary` before this provider sees the text.\n *   - The audio URL returned by DashScope is hosted on Alibaba's CDN (variable\n *     hostname). We do not SSRF-guard the audio fetch because:\n *       (a) the URL is provided by the same provider we already trust for the\n *           initial synthesis call,\n *       (b) it's HTTPS-only and short-lived (~5 min TTL),\n *       (c) maintaining a hostname allowlist breaks every time Alibaba rotates\n *           CDN domains.\n *     A dedicated `assertSafeUrl` for trusted-CDN responses can be added later\n *     as a hardening pass.\n */\n\nimport {\n  ProviderHttpError,\n  fetchWithTimeoutGuarded,\n  postJsonRequest,\n} from '../../../media-shared/http/index.js';\nimport { createLogger } from '../../../utils/logger.js';\nimport { registerSpeechProvider } from '../speech-registry.js';\nimport type {\n  SpeechDirectiveTokenParseContext,\n  SpeechDirectiveTokenParseResult,\n  SpeechProviderConfig,\n  SpeechProviderConfiguredContext,\n  SpeechProviderPlugin,\n  SpeechProviderResolveConfigContext,\n  SpeechSynthesisRequest,\n  SpeechSynthesisResult,\n} from '../speech-provider-types.js';\n\nconst log = createLogger('SpeechProvider:Alibaba');\n\nconst DEFAULT_BASE_URL =\n  'https://dashscope.aliyuncs.com/api/v1/services/aigc/multimodal-generation/generation';\nconst DEFAULT_MODEL = 'qwen-tts';\nconst DEFAULT_VOICE = 'longxiaochun';\nconst ENV_KEY = 'DASHSCOPE_API_KEY';\nconst MAX_TEXT_LENGTH = 512;\nconst ALIBABA_VOICES = [\n  'Cherry',\n  'Ethan',\n  'Serena',\n  'Chelsie',\n  'longxiaochun',\n  'longxiaobai',\n  'longwan',\n  'longcheng',\n] as const;\n\ninterface AlibabaTtsConfig extends Record<string, unknown> {\n  apiKey?: string;\n  baseUrl: string;\n  model: string;\n  voice: string;\n}\n\ninterface CosyVoiceResponse {\n  output: {\n    audio?: { url?: string; data?: string };\n    speech?: string;\n    speech_url?: string;\n    finish_reason?: string;\n  };\n  usage?: { characters?: number };\n  request_id?: string;\n}\n\nfunction asObject(value: unknown): Record<string, unknown> | undefined {\n  return typeof value === 'object' && value !== null && !Array.isArray(value)\n    ? (value as Record<string, unknown>)\n    : undefined;\n}\n\nfunction trimToUndefined(value: unknown): string | undefined {\n  if (typeof value !== 'string') {\n    return undefined;\n  }\n  const trimmed = value.trim();\n  return trimmed.length > 0 ? trimmed : undefined;\n}\n\nfunction normalizeConfig(rawConfig: Record<string, unknown>): AlibabaTtsConfig {\n  const raw = asObject(rawConfig.alibaba) ?? {};\n  return {\n    apiKey: trimToUndefined(raw.apiKey),\n    baseUrl: trimToUndefined(raw.baseUrl) ?? DEFAULT_BASE_URL,\n    model: trimToUndefined(raw.model ?? raw.modelId) ?? DEFAULT_MODEL,\n    voice: trimToUndefined(raw.voice ?? raw.voiceId) ?? DEFAULT_VOICE,\n  };\n}\n\nfunction readProviderConfig(config: SpeechProviderConfig): AlibabaTtsConfig {\n  return {\n    apiKey: trimToUndefined(config.apiKey),\n    baseUrl: trimToUndefined(config.baseUrl) ?? DEFAULT_BASE_URL,\n    model: trimToUndefined(config.model ?? config.modelId) ?? DEFAULT_MODEL,\n    voice: trimToUndefined(config.voice ?? config.voiceId) ?? DEFAULT_VOICE,\n  };\n}\n\nfunction resolveApiKey(config: AlibabaTtsConfig): string | undefined {\n  return config.apiKey ?? trimToUndefined(process.env[ENV_KEY]);\n}\n\nfunction parseDirectiveTokenInternal(\n  ctx: SpeechDirectiveTokenParseContext,\n): SpeechDirectiveTokenParseResult {\n  switch (ctx.key) {\n    case 'voice':\n    case 'voice_id':\n    case 'voiceid':\n    case 'alibaba_voice':\n    case 'alibabavoice':\n      if (!ctx.policy.allowVoice) {\n        return { handled: true };\n      }\n      return { handled: true, overrides: { voice: ctx.value } };\n    case 'model':\n    case 'model_id':\n    case 'modelid':\n    case 'alibaba_model':\n    case 'alibabamodel':\n      if (!ctx.policy.allowModelId) {\n        return { handled: true };\n      }\n      return { handled: true, overrides: { model: ctx.value } };\n    default:\n      return { handled: false };\n  }\n}\n\nexport const alibabaSpeechProvider: SpeechProviderPlugin = {\n  id: 'alibaba',\n  aliases: ['dashscope', 'qwen-tts'],\n  autoSelectOrder: 25,\n\n  resolveConfig: (ctx: SpeechProviderResolveConfigContext) => normalizeConfig(ctx.rawConfig),\n\n  isConfigured: (ctx: SpeechProviderConfiguredContext) =>\n    Boolean(resolveApiKey(readProviderConfig(ctx.providerConfig))),\n\n  parseDirectiveToken: parseDirectiveTokenInternal,\n\n  listVoices: async () => ALIBABA_VOICES.map((id) => ({ id, name: id })),\n\n  synthesize: async (req: SpeechSynthesisRequest): Promise<SpeechSynthesisResult> => {\n    const config = readProviderConfig(req.providerConfig);\n    const apiKey = resolveApiKey(config);\n    if (!apiKey) {\n      throw new Error(\n        `Alibaba TTS API key missing (set ${ENV_KEY} or messages.tts.providers.alibaba.apiKey)`,\n      );\n    }\n    if (req.text.length > MAX_TEXT_LENGTH) {\n      throw new Error(\n        `Alibaba TTS text exceeds ${MAX_TEXT_LENGTH} char limit (got ${req.text.length})`,\n      );\n    }\n\n    const overrides = req.providerOverrides ?? {};\n    const model = trimToUndefined(overrides.model ?? overrides.modelId) ?? config.model;\n    const voice = trimToUndefined(overrides.voice ?? overrides.voiceId) ?? config.voice;\n\n    log.debug({ model, voice, textLength: req.text.length }, 'Calling Alibaba TTS');\n\n    const response = await postJsonRequest(config.baseUrl, {\n      timeoutMs: req.timeoutMs,\n      label: 'Alibaba TTS',\n      headers: {\n        Authorization: `Bearer ${apiKey}`,\n        'X-DashScope-DataInspection': 'disable',\n      },\n      body: {\n        model,\n        input: { text: req.text },\n        parameters: { voice },\n      },\n    });\n\n    const data = (await response.json()) as CosyVoiceResponse;\n\n    // DashScope occasionally returns `finish_reason: \"null\"` (string) on failure.\n    if (data.output?.finish_reason === 'null' && !data.output?.audio?.url) {\n      throw new Error(`Alibaba TTS API error: ${JSON.stringify(data)}`);\n    }\n\n    let audioBuffer: Buffer;\n    const audioUrl = data.output?.audio?.url ?? data.output?.speech_url;\n    if (audioUrl) {\n      // CDN URL — see file-level DECISION on why we skip SSRF here.\n      const audioResponse = await fetchWithTimeoutGuarded(audioUrl, {\n        timeoutMs: req.timeoutMs,\n        label: 'Alibaba TTS audio download',\n        allowPrivateNetwork: false,\n      });\n      if (!audioResponse.ok) {\n        throw new ProviderHttpError({\n          label: 'Alibaba TTS audio download',\n          status: audioResponse.status,\n          detail: audioResponse.statusText,\n        });\n      }\n      audioBuffer = Buffer.from(await audioResponse.arrayBuffer());\n    } else if (data.output?.speech) {\n      audioBuffer = Buffer.from(data.output.speech, 'base64');\n    } else if (data.output?.audio?.data) {\n      audioBuffer = Buffer.from(data.output.audio.data, 'base64');\n    } else {\n      throw new Error('No audio returned from Alibaba TTS');\n    }\n\n    log.debug(\n      { size: audioBuffer.length, characters: data.usage?.characters, requestId: data.request_id },\n      'Alibaba TTS completed',\n    );\n\n    return {\n      audioBuffer,\n      outputFormat: 'wav',\n      fileExtension: 'wav',\n      // wav is NOT a Telegram voice-note format → ffmpeg compresses downstream.\n      voiceCompatible: false,\n    };\n  },\n\n  // synthesizeStream intentionally omitted — see file-level DECISION.\n};\n\nregisterSpeechProvider(alibabaSpeechProvider);\n"],"mappings":";;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;aA+BwD;AAaxD,MAAM,MAAM,aAAa,yBAAyB;AAElD,MAAM,mBACJ;AACF,MAAM,gBAAgB;AACtB,MAAM,gBAAgB;AACtB,MAAM,UAAU;AAChB,MAAM,kBAAkB;AACxB,MAAM,iBAAiB;CACrB;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACD;AAoBD,SAAS,SAAS,OAAqD;AACrE,QAAO,OAAO,UAAU,YAAY,UAAU,QAAQ,CAAC,MAAM,QAAQ,MAAM,GACtE,QACD,KAAA;;AAGN,SAAS,gBAAgB,OAAoC;AAC3D,KAAI,OAAO,UAAU,SACnB;CAEF,MAAM,UAAU,MAAM,MAAM;AAC5B,QAAO,QAAQ,SAAS,IAAI,UAAU,KAAA;;AAGxC,SAAS,gBAAgB,WAAsD;CAC7E,MAAM,MAAM,SAAS,UAAU,QAAQ,IAAI,EAAE;AAC7C,QAAO;EACL,QAAQ,gBAAgB,IAAI,OAAO;EACnC,SAAS,gBAAgB,IAAI,QAAQ,IAAI;EACzC,OAAO,gBAAgB,IAAI,SAAS,IAAI,QAAQ,IAAI;EACpD,OAAO,gBAAgB,IAAI,SAAS,IAAI,QAAQ,IAAI;EACrD;;AAGH,SAAS,mBAAmB,QAAgD;AAC1E,QAAO;EACL,QAAQ,gBAAgB,OAAO,OAAO;EACtC,SAAS,gBAAgB,OAAO,QAAQ,IAAI;EAC5C,OAAO,gBAAgB,OAAO,SAAS,OAAO,QAAQ,IAAI;EAC1D,OAAO,gBAAgB,OAAO,SAAS,OAAO,QAAQ,IAAI;EAC3D;;AAGH,SAAS,cAAc,QAA8C;AACnE,QAAO,OAAO,UAAU,gBAAgB,QAAQ,IAAI,SAAS;;AAG/D,SAAS,4BACP,KACiC;AACjC,SAAQ,IAAI,KAAZ;EACE,KAAK;EACL,KAAK;EACL,KAAK;EACL,KAAK;EACL,KAAK;AACH,OAAI,CAAC,IAAI,OAAO,WACd,QAAO,EAAE,SAAS,MAAM;AAE1B,UAAO;IAAE,SAAS;IAAM,WAAW,EAAE,OAAO,IAAI,OAAO;IAAE;EAC3D,KAAK;EACL,KAAK;EACL,KAAK;EACL,KAAK;EACL,KAAK;AACH,OAAI,CAAC,IAAI,OAAO,aACd,QAAO,EAAE,SAAS,MAAM;AAE1B,UAAO;IAAE,SAAS;IAAM,WAAW,EAAE,OAAO,IAAI,OAAO;IAAE;EAC3D,QACE,QAAO,EAAE,SAAS,OAAO;;;AAI/B,MAAa,wBAA8C;CACzD,IAAI;CACJ,SAAS,CAAC,aAAa,WAAW;CAClC,iBAAiB;CAEjB,gBAAgB,QAA4C,gBAAgB,IAAI,UAAU;CAE1F,eAAe,QACb,QAAQ,cAAc,mBAAmB,IAAI,eAAe,CAAC,CAAC;CAEhE,qBAAqB;CAErB,YAAY,YAAY,eAAe,KAAK,QAAQ;EAAE;EAAI,MAAM;EAAI,EAAE;CAEtE,YAAY,OAAO,QAAgE;EACjF,MAAM,SAAS,mBAAmB,IAAI,eAAe;EACrD,MAAM,SAAS,cAAc,OAAO;AACpC,MAAI,CAAC,OACH,OAAM,IAAI,MACR,oCAAoC,QAAQ,4CAC7C;AAEH,MAAI,IAAI,KAAK,SAAS,gBACpB,OAAM,IAAI,MACR,4BAA4B,gBAAgB,mBAAmB,IAAI,KAAK,OAAO,GAChF;EAGH,MAAM,YAAY,IAAI,qBAAqB,EAAE;EAC7C,MAAM,QAAQ,gBAAgB,UAAU,SAAS,UAAU,QAAQ,IAAI,OAAO;EAC9E,MAAM,QAAQ,gBAAgB,UAAU,SAAS,UAAU,QAAQ,IAAI,OAAO;AAE9E,MAAI,MAAM;GAAE;GAAO;GAAO,YAAY,IAAI,KAAK;GAAQ,EAAE,sBAAsB;EAgB/E,MAAM,OAAQ,OAAM,MAdG,gBAAgB,OAAO,SAAS;GACrD,WAAW,IAAI;GACf,OAAO;GACP,SAAS;IACP,eAAe,UAAU;IACzB,8BAA8B;IAC/B;GACD,MAAM;IACJ;IACA,OAAO,EAAE,MAAM,IAAI,MAAM;IACzB,YAAY,EAAE,OAAO;IACtB;GACF,CAAC,EAE2B,MAAM;AAGnC,MAAI,KAAK,QAAQ,kBAAkB,UAAU,CAAC,KAAK,QAAQ,OAAO,IAChE,OAAM,IAAI,MAAM,0BAA0B,KAAK,UAAU,KAAK,GAAG;EAGnE,IAAI;EACJ,MAAM,WAAW,KAAK,QAAQ,OAAO,OAAO,KAAK,QAAQ;AACzD,MAAI,UAAU;GAEZ,MAAM,gBAAgB,MAAM,wBAAwB,UAAU;IAC5D,WAAW,IAAI;IACf,OAAO;IACP,qBAAqB;IACtB,CAAC;AACF,OAAI,CAAC,cAAc,GACjB,OAAM,IAAI,kBAAkB;IAC1B,OAAO;IACP,QAAQ,cAAc;IACtB,QAAQ,cAAc;IACvB,CAAC;AAEJ,iBAAc,OAAO,KAAK,MAAM,cAAc,aAAa,CAAC;aACnD,KAAK,QAAQ,OACtB,eAAc,OAAO,KAAK,KAAK,OAAO,QAAQ,SAAS;WAC9C,KAAK,QAAQ,OAAO,KAC7B,eAAc,OAAO,KAAK,KAAK,OAAO,MAAM,MAAM,SAAS;MAE3D,OAAM,IAAI,MAAM,qCAAqC;AAGvD,MAAI,MACF;GAAE,MAAM,YAAY;GAAQ,YAAY,KAAK,OAAO;GAAY,WAAW,KAAK;GAAY,EAC5F,wBACD;AAED,SAAO;GACL;GACA,cAAc;GACd,eAAe;GAEf,iBAAiB;GAClB;;CAIJ;AAED,uBAAuB,sBAAsB"}

package/dist/src/voice/tts/providers/edge-speech.js

@@ -1,9 +1,9 @@
 import { createLogger } from "../../../utils/logger/index.js";
 import { init_logger } from "../../../utils/logger.js";
 import { registerSpeechProvider } from "../speech-registry.js";
-import { tmpdir } from "node:os";
-import path from "node:path";
 import { mkdtempSync, readFileSync, rmSync } from "node:fs";
+import path from "node:path";
+import { tmpdir } from "node:os";
 //#region src/voice/tts/providers/edge-speech.ts
 /**
 * Edge TTS provider — wraps the `node-edge-tts` package. Requires no API key

package/dist/src/voice/tts/types.d.ts

@@ -72,36 +72,8 @@ export interface TTSConfig {
     summarization?: TTSSummarizationConfig;
     /** Allow model to override TTS parameters */
     modelOverrides?: TTSModelOverrideConfig;
-    /** OpenClaw-aligned provider settings map (`messages.tts.providers.<id>`). */
+    /** Provider settings map (`messages.tts.providers.<id>`). */
     providers?: Record<string, Record<string, unknown>>;
-    alibaba?: {
-        apiKey?: string;
-        model?: string;
-        voice?: string;
-    };
-    openai?: {
-        apiKey?: string;
-        model?: string;
-        voice?: string;
-    };
-    edge?: {
-        enabled?: boolean;
-        voice?: string;
-        lang?: string;
-        outputFormat?: string;
-        pitch?: string;
-        rate?: string;
-        volume?: string;
-        proxy?: string;
-        timeoutMs?: number;
-    };
-    minimax?: {
-        apiKey?: string;
-        model?: string;
-        voice?: string;
-    };
-    /** Extension / legacy flat provider config (e.g. `tts-local-cli`). */
-    [providerId: string]: unknown;
 }
 export interface TTSSummarizationConfig {
     /** When true (default), long text is summarized via LLM before TTS */

package/dist/src/voice/tts/types.js

@@ -25,23 +25,25 @@ const DEFAULT_TTS_CONFIG = {
 		allowNormalization: false,
 		allowSeed: false
 	},
-	alibaba: {
-		model: "qwen-tts",
-		voice: "Cherry"
-	},
-	openai: {
-		model: "tts-1",
-		voice: "alloy"
-	},
-	edge: {
-		enabled: true,
-		voice: "en-US-MichelleNeural",
-		lang: "en-US",
-		outputFormat: "audio-24khz-48kbitrate-mono-mp3"
-	},
-	minimax: {
-		model: "speech-2.8-hd",
-		voice: "male-qn-qingse"
+	providers: {
+		alibaba: {
+			model: "qwen-tts",
+			voice: "Cherry"
+		},
+		openai: {
+			model: "tts-1",
+			voice: "alloy"
+		},
+		edge: {
+			enabled: true,
+			voice: "en-US-MichelleNeural",
+			lang: "en-US",
+			outputFormat: "audio-24khz-48kbitrate-mono-mp3"
+		},
+		minimax: {
+			model: "speech-2.8-hd",
+			voice: "male-qn-qingse"
+		}
 	}
 };
 //#endregion