Update app.py with gr.Audio()

What changed:
Replaced gr.Video() with gr.Audio() — better UX for audio files, still accepts video via "upload" source
Removed SRT generation entirely (you only wanted text output)
Simplified return values to just (text, txt_path)
Updated UI labels to emphasize "copy & paste" workflow
Kept all the core logic identical — ffmpeg handles both audio and video extraction seamlessly

app.py

@@ -35,7 +35,7 @@ def extract_audio(video_path: str, workdir: str) -> str:
     cmd = [
         "ffmpeg", "-y",
         "-i", video_path,
-        "-vn",               # drop video stream
+        "-vn",               # drop video stream (safe for pure audio too)
         "-acodec", "pcm_s16le",
         "-ar", "16000",
         "-ac", "1",
@@ -79,20 +79,20 @@ def safe_stem(filename: str) -> str:
 # ---------------------------------------------------------------------------
 # Main transcription handler
 # ---------------------------------------------------------------------------
-def transcribe(video_file, language, task, progress=gr.Progress(track_tqdm=True)):
-    if video_file is None:
-        return "Please upload a video or audio file first.", None, None
+def transcribe(media_file, language, task, progress=gr.Progress(track_tqdm=True)):
+    if media_file is None:
+        return "", ""
 
-    src_path = video_file if isinstance(video_file, str) else video_file.name
+    src_path = media_file if isinstance(media_file, str) else media_file.name
     ext = Path(src_path).suffix.lower()
     if ext and ext not in SUPPORTED:
-        return f"Unsupported file type: `{ext}`", None, None
+        return f"Unsupported file type: `{ext}`", ""
 
     with tempfile.TemporaryDirectory() as workdir:
-        progress(0.1, desc="Extracting audio with ffmpeg...")
+        progress(0.1, desc="Extracting audio...")
         audio_path = extract_audio(src_path, workdir)
 
-        progress(0.3, desc="Transcribing with Whisper (this can take a while)...")
+        progress(0.3, desc="Transcribing (this can take a while)...")
         kwargs = {"generate_kwargs": {"task": task}}
         if language != "auto":
             kwargs["generate_kwargs"]["language"] = language
@@ -101,24 +101,13 @@ def transcribe(video_file, language, task, progress=gr.Progress(track_tqdm=True)
         text = result["text"].strip()
         chunks = result.get("chunks") or []
 
-        progress(0.9, desc="Preparing download files...")
+        progress(0.9, desc="Finalizing...")
         stem = safe_stem(src_path)
-
         txt_path = os.path.join(workdir, f"{stem}_transcript.txt")
         with open(txt_path, "w", encoding="utf-8") as f:
             f.write(text + "\n")
-        final_txt = os.path.join(tempfile.gettempdir(), f"{stem}_transcript.txt")
-        shutil.copy(txt_path, final_txt)
-
-        final_srt = None
-        if chunks:
-            srt_path = os.path.join(workdir, f"{stem}_subtitles.srt")
-            with open(srt_path, "w", encoding="utf-8") as f:
-                f.write(build_srt(chunks))
-            final_srt = os.path.join(tempfile.gettempdir(), f"{stem}_subtitles.srt")
-            shutil.copy(srt_path, final_srt)
 
-    return text, final_txt, final_srt
+    return text, txt_path
 
 
 # ---------------------------------------------------------------------------
@@ -130,24 +119,21 @@ LANGUAGES = [
     "turkish", "polish", "swedish", "ukrainian", "vietnamese", "indonesian",
 ]
 
-with gr.Blocks(title="Video Transcriber") as demo:
+with gr.Blocks(title="Transcriber") as demo:
     gr.Markdown(
         """
-# 🎬 → 📝 Video Transcriber (Speech-to-Text)
-
-Upload a video (or audio) file and get a full text transcription powered by
-**OpenAI Whisper** running on 🤗 Transformers.
-
+# 🎙️ Audio/Video Transcriber
+Upload any audio or video file and get a clean text transcription powered by **OpenAI Whisper**.
 **How it works**
-1. **Upload** a video file (`.mp4`, `.mov`, `.mkv`, `.webm`, ...) — audio files work too.
-2. Click **Transcribe**. The audio track is extracted with `ffmpeg` and transcribed.
-3. **Read** the transcript on the right and **download** it as `.txt` or timed subtitles `.srt`.
+1. **Upload** an audio (.mp3, .wav, .m4a...) or video (.mp4, .mov, .mkv...) file.
+2. Click **Transcribe**. The audio is extracted and transcribed automatically.
+3. **Copy & paste** the transcript below or download it as a `.txt` file.
 """
     )
 
     with gr.Row():
         with gr.Column(scale=1):
-            video_in = gr.Video(label="Upload video (or drop an audio file)", sources=["upload"])
+            media_in = gr.Audio(label="Audio/Video", sources=["upload"])
             language = gr.Dropdown(
                 choices=LANGUAGES, value="auto",
                 label="Language (optional — auto-detect by default)",
@@ -157,16 +143,15 @@ Upload a video (or audio) file and get a full text transcription powered by
                 label="Task ('translate' translates speech into English)",
             )
             run_btn = gr.Button("🚀 Transcribe", variant="primary")
+
         with gr.Column(scale=1):
-            transcript_out = gr.Textbox(label="Transcript", lines=22)
-            with gr.Row():
-                txt_file = gr.File(label="⬇️ Download transcript (.txt)")
-                srt_file = gr.File(label="⬇️ Download subtitles (.srt)")
+            transcript_out = gr.Textbox(label="Transcript (copy & paste)", lines=25)
+            txt_file = gr.File(label="⬇️ Download .txt")
 
     run_btn.click(
         fn=transcribe,
-        inputs=[video_in, language, task],
-        outputs=[transcript_out, txt_file, srt_file],
+        inputs=[media_in, language, task],
+        outputs=[transcript_out, txt_file],
     )
 
 if __name__ == "__main__":