diff --git a/src/mcspeak/audio.py b/src/mcspeak/audio.py
index 4574649..6c66d50 100644
--- a/src/mcspeak/audio.py
+++ b/src/mcspeak/audio.py
@@ -331,7 +331,8 @@ async def record_audio_until_silence(
 
     if not pcm_buf:
         raise PlaybackError(
-            f"VAD recording captured no audio. Check source '{source or 'default'}'."
+            f"No audio captured — is a microphone connected? "
+            f"(source '{source or 'default'}' produced no samples)"
         )
 
     # Write the accumulated PCM as a WAV file.
diff --git a/src/mcspeak/server.py b/src/mcspeak/server.py
index a1fc7c0..5bc9bb2 100644
--- a/src/mcspeak/server.py
+++ b/src/mcspeak/server.py
@@ -895,7 +895,10 @@ async def listen(
                 warmup_ms=warmup_ms,
             )
         except Exception as e:
-            return {"error": f"VAD recording failed: {e}"}
+            # Include the exception TYPE — some (e.g. asyncio.TimeoutError)
+            # stringify to "", which otherwise hides a missing-mic failure.
+            detail = str(e) or "no detail (often a disconnected/blocked mic)"
+            return {"error": f"VAD recording failed ({type(e).__name__}): {detail}"}
     else:
         await ctx.info(f"Listening for {duration_seconds:.1f}s (source={source or 'default'})...")
         try:
@@ -904,7 +907,8 @@ async def listen(
                 ready_tone=start_tone, warmup_ms=warmup_ms,
             )
         except Exception as e:
-            return {"error": f"Recording failed: {e}"}
+            detail = str(e) or "no detail (often a disconnected/blocked mic)"
+            return {"error": f"Recording failed ({type(e).__name__}): {detail}"}
 
     # "Capture done" tone AFTER pw-record stops. Audible "over and out"
     # so the user knows they can stop talking and the system is now