feat(backend,frontend,mcp): any voice or round video can be transcribed on request, media is found by its bytes

frontend/src/lib/api/endpoints.ts

@@ -575,15 +575,23 @@ export function fetchMedia(
   });
 }
 
+export interface TranscribeResult {
+  error: string | null;
+  status: "done" | "failed" | "pending";
+  text: string | null;
+}
+
 export function transcribeMedia(
   chatId: number,
-  messageId: number
-): Promise<{ job_id: number }> {
-  return request<{ job_id: number }>("/media/transcribe", {
+  messageId: number,
+  wait: number
+): Promise<TranscribeResult> {
+  return request<TranscribeResult>("/media/transcribe", {
     body: {
       account_id: accounts.selectedId,
       chat_id: chatId,
       message_id: messageId,
+      wait,
     },
     method: "POST",
   });