TTS: add provider personas

src/plugin-sdk/config-runtime.ts

@@ -133,10 +133,15 @@ export type {
   TelegramInlineButtonsScope,
   TelegramNetworkConfig,
   TelegramTopicConfig,
+  ResolvedTtsPersona,
   TtsAutoMode,
   TtsConfig,
   TtsMode,
   TtsModelOverrideConfig,
+  TtsPersonaConfig,
+  TtsPersonaFallbackPolicy,
+  TtsPersonaPromptConfig,
+  TtsPersonaRewriteConfig,
   TtsProvider,
 } from "../config/types.js";
 export {

src/plugin-sdk/speech-core.ts

@@ -9,11 +9,14 @@ export type {
   SpeechModelOverridePolicy,
   SpeechProviderConfig,
   SpeechProviderConfiguredContext,
+  SpeechProviderPreparedSynthesis,
+  SpeechProviderPrepareSynthesisContext,
   SpeechProviderResolveConfigContext,
   SpeechProviderResolveTalkConfigContext,
   SpeechProviderResolveTalkOverridesContext,
   SpeechProviderOverrides,
   SpeechSynthesisRequest,
+  SpeechSynthesisTarget,
   SpeechTelephonySynthesisRequest,
   SpeechVoiceOption,
   TtsDirectiveOverrides,
@@ -35,6 +38,7 @@ export {
   listSpeechProviders,
   normalizeSpeechProviderId,
 } from "../tts/provider-registry.js";
+export { resolveEffectiveTtsConfig } from "../tts/tts-config.js";
 export { normalizeTtsAutoMode, TTS_AUTO_MODES } from "../tts/tts-auto-mode.js";
 export {
   asBoolean,

src/plugin-sdk/speech.ts

@@ -12,11 +12,14 @@ export type {
   SpeechModelOverridePolicy,
   SpeechProviderConfig,
   SpeechProviderConfiguredContext,
+  SpeechProviderPreparedSynthesis,
+  SpeechProviderPrepareSynthesisContext,
   SpeechProviderResolveConfigContext,
   SpeechProviderResolveTalkConfigContext,
   SpeechProviderResolveTalkOverridesContext,
   SpeechProviderOverrides,
   SpeechSynthesisRequest,
+  SpeechSynthesisTarget,
   SpeechTelephonySynthesisRequest,
   SpeechVoiceOption,
   TtsDirectiveOverrides,

src/plugin-sdk/tts-runtime.ts

@@ -40,6 +40,10 @@ export const getTtsMaxLength: FacadeModule["getTtsMaxLength"] = createLazyFacade
   loadFacadeModule,
   "getTtsMaxLength",
 );
+export const getTtsPersona: FacadeModule["getTtsPersona"] = createLazyFacadeRuntimeValue(
+  loadFacadeModule,
+  "getTtsPersona",
+);
 export const getTtsProvider: FacadeModule["getTtsProvider"] = createLazyFacadeRuntimeValue(
   loadFacadeModule,
   "getTtsProvider",
@@ -56,6 +60,10 @@ export const listSpeechVoices: FacadeModule["listSpeechVoices"] = createLazyFaca
   loadFacadeModule,
   "listSpeechVoices",
 );
+export const listTtsPersonas: FacadeModule["listTtsPersonas"] = createLazyFacadeRuntimeValue(
+  loadFacadeModule,
+  "listTtsPersonas",
+);
 export const maybeApplyTtsToPayload: FacadeModule["maybeApplyTtsToPayload"] =
   createLazyFacadeRuntimeValue(loadFacadeModule, "maybeApplyTtsToPayload");
 export const resolveExplicitTtsOverrides: FacadeModule["resolveExplicitTtsOverrides"] =
@@ -90,6 +98,10 @@ export const setTtsMaxLength: FacadeModule["setTtsMaxLength"] = createLazyFacade
   loadFacadeModule,
   "setTtsMaxLength",
 );
+export const setTtsPersona: FacadeModule["setTtsPersona"] = createLazyFacadeRuntimeValue(
+  loadFacadeModule,
+  "setTtsPersona",
+);
 export const setTtsProvider: FacadeModule["setTtsProvider"] = createLazyFacadeRuntimeValue(
   loadFacadeModule,
   "setTtsProvider",

src/plugin-sdk/tts-runtime.types.ts

@@ -1,5 +1,5 @@
 import type { OpenClawConfig } from "../config/types.openclaw.js";
-import type { TtsAutoMode, TtsProvider } from "../config/types.tts.js";
+import type { ResolvedTtsPersona, TtsAutoMode, TtsProvider } from "../config/types.tts.js";
 import type {
   SpeechProviderConfig,
   SpeechVoiceOption,
@@ -24,6 +24,8 @@ export type TtsProviderAttempt = {
   provider: string;
   outcome: "success" | "skipped" | "failed";
   reasonCode: TtsAttemptReasonCode;
+  persona?: string;
+  personaBinding?: "applied" | "missing" | "none";
   latencyMs?: number;
   error?: string;
 };
@@ -34,6 +36,7 @@ export type TtsStatusEntry = {
   textLength: number;
   summarized: boolean;
   provider?: string;
+  persona?: string;
   fallbackFrom?: string;
   attemptedProviders?: string[];
   attempts?: TtsProviderAttempt[];
@@ -126,6 +129,7 @@ export type TtsResult = {
   error?: string;
   latencyMs?: number;
   provider?: string;
+  persona?: string;
   fallbackFrom?: string;
   attemptedProviders?: string[];
   attempts?: TtsProviderAttempt[];
@@ -141,6 +145,7 @@ export type TtsSynthesisResult = {
   error?: string;
   latencyMs?: number;
   provider?: string;
+  persona?: string;
   fallbackFrom?: string;
   attemptedProviders?: string[];
   attempts?: TtsProviderAttempt[];
@@ -156,6 +161,7 @@ export type TtsTelephonyResult = {
   error?: string;
   latencyMs?: number;
   provider?: string;
+  persona?: string;
   fallbackFrom?: string;
   attemptedProviders?: string[];
   attempts?: TtsProviderAttempt[];
@@ -179,6 +185,7 @@ export type TtsRuntimeFacade = {
     cfg?: OpenClawConfig,
   ) => SpeechProviderConfig;
   getTtsMaxLength: (prefsPath: string) => number;
+  getTtsPersona: (config: ResolvedTtsConfig, prefsPath: string) => ResolvedTtsPersona | undefined;
   getTtsProvider: (config: ResolvedTtsConfig, prefsPath: string) => TtsProvider;
   isSummarizationEnabled: (prefsPath: string) => boolean;
   isTtsEnabled: (config: ResolvedTtsConfig, prefsPath: string, sessionAuto?: string) => boolean;
@@ -188,6 +195,7 @@ export type TtsRuntimeFacade = {
     cfg?: OpenClawConfig,
   ) => boolean;
   listSpeechVoices: ListSpeechVoices;
+  listTtsPersonas: (config: ResolvedTtsConfig) => ResolvedTtsPersona[];
   maybeApplyTtsToPayload: (params: MaybeApplyTtsToPayloadParams) => Promise<ReplyPayload>;
   resolveExplicitTtsOverrides: (params: ResolveExplicitTtsOverridesParams) => TtsDirectiveOverrides;
   resolveTtsAutoMode: (params: ResolveTtsAutoModeParams) => TtsAutoMode;
@@ -199,6 +207,7 @@ export type TtsRuntimeFacade = {
   setTtsAutoMode: (prefsPath: string, mode: TtsAutoMode) => void;
   setTtsEnabled: (prefsPath: string, enabled: boolean) => void;
   setTtsMaxLength: (prefsPath: string, maxLength: number) => void;
+  setTtsPersona: (prefsPath: string, persona: string | null | undefined) => void;
   setTtsProvider: (prefsPath: string, provider: TtsProvider) => void;
   synthesizeSpeech: (params: TtsRequestParams) => Promise<TtsSynthesisResult>;
   textToSpeech: TextToSpeech;