diff --git a/extensions/tts-local-cli/speech-provider.test.ts b/extensions/tts-local-cli/speech-provider.test.ts
index 75b09ee8f1c..b0d9ca23f26 100644
--- a/extensions/tts-local-cli/speech-provider.test.ts
+++ b/extensions/tts-local-cli/speech-provider.test.ts
@@ -80,7 +80,18 @@ describe("buildCliSpeechProvider", () => {
       if (typeof outputPath !== "string") {
         throw new Error("missing ffmpeg output path");
       }
-      writeFileSync(outputPath, Buffer.from(`converted:${path.extname(outputPath)}`));
+      const forcedFormatIndex = args.lastIndexOf("-f");
+      const forcedFormat =
+        forcedFormatIndex >= 0 && typeof args[forcedFormatIndex + 1] === "string"
+          ? args[forcedFormatIndex + 1]
+          : undefined;
+      const extension =
+        forcedFormat === "s16le"
+          ? ".pcm"
+          : forcedFormat
+            ? `.${forcedFormat}`
+            : path.extname(outputPath);
+      writeFileSync(outputPath, Buffer.from(`converted:${extension}`));
     });
   });
 
diff --git a/extensions/tts-local-cli/speech-provider.ts b/extensions/tts-local-cli/speech-provider.ts
index aece764806c..283057ce748 100644
--- a/extensions/tts-local-cli/speech-provider.ts
+++ b/extensions/tts-local-cli/speech-provider.ts
@@ -275,11 +275,11 @@ async function convertAudio(
   const outputPath = path.join(outputDir, outputFileName);
   const args = ["-y", "-i", inputPath];
   if (target === "opus") {
-    args.push("-c:a", "libopus", "-b:a", "64k");
+    args.push("-c:a", "libopus", "-b:a", "64k", "-f", "opus");
   } else if (target === "wav") {
-    args.push("-c:a", "pcm_s16le");
+    args.push("-c:a", "pcm_s16le", "-f", "wav");
   } else {
-    args.push("-c:a", "libmp3lame", "-b:a", "128k");
+    args.push("-c:a", "libmp3lame", "-b:a", "128k", "-f", "mp3");
   }
   await writeExternalFileWithinRoot({
     rootDir: outputDir,