feat: implement tier-a4 tts voice output replies

src/config/schema.ts

@@ -730,6 +730,24 @@ const audioSchema = z.object({
   talk_mode: talkModeSchema,
 }).default({});
 
+const ttsOutputFormatSchema = z.enum(['mp3', 'wav', 'opus']);
+
+const ttsProviderSchema = z.object({
+  type: z.enum(['openai', 'custom']).default('openai'),
+  endpoint: z.string().optional(),
+  api_key: z.string().optional(),
+  model: z.string().default('gpt-4o-mini-tts'),
+  voice: z.string().default('alloy'),
+  format: ttsOutputFormatSchema.default('mp3'),
+});
+
+const ttsSchema = z.object({
+  enabled: z.boolean().default(false),
+  /** Restrict voice replies to selected channels. Empty means all channels. */
+  enabled_channels: z.array(z.string().min(1)).default([]),
+  provider: ttsProviderSchema.optional(),
+}).default({});
+
 // ── Tool policy schemas ──────────────────────────────────────────────
 
 const toolProfileEnum = z.enum(['minimal', 'messaging', 'coding', 'full']);
@@ -912,6 +930,7 @@ export const configSchema = z.object({
   retry: retrySchema,
   web_search: webSearchSchema,
   audio: audioSchema,
+  tts: ttsSchema,
   prompt: promptSchema,
   tools: toolsSchema,
   sandbox: sandboxSchema,
@@ -936,6 +955,7 @@ export type CompactionConfig = z.infer<typeof compactionSchema>;
 export type MemoryConfig = z.infer<typeof memorySchema>;
 export type WebSearchConfig = z.infer<typeof webSearchSchema>;
 export type AudioConfig = z.infer<typeof audioSchema>;
+export type TtsConfig = z.infer<typeof ttsSchema>;
 export type ProcessConfig = z.infer<typeof processSchema>;
 export type BrowserConfig = z.infer<typeof browserSchema>;
 export type K8sConfig = z.infer<typeof k8sSchema>;