transcriber word count app file

app.py

@@ -0,0 +1,43 @@
+import modal
+
+app = modal.App("whisper-agentic")
+
+# Define the Modal image environment
+image = (
+    modal.Image.debian_slim().apt_install("ffmpeg").pip_install("transformers", "torch", "gradio", "torchaudio", "accelerate", "optimum[diffusers]", "gradio_client")
+)
+
+@app.function(image=image, gpu="A10G", timeout=7200)
+def run_gradio():
+    from transformers import pipeline
+    import gradio as gr
+
+    # Load Whisper pipeline
+    transcriber = pipeline(model="openai/whisper-large-v2", return_timestamps=True)
+
+    # Agentic function
+    def transcribe_and_analyze(audio):
+        result = transcriber(audio)
+        transcript = result["text"]
+
+        word_count = len(transcript.split())
+        char_count = len(transcript.replace(" ", ""))
+
+        return {
+            "Transcript": transcript,
+            "Word Count": word_count,
+            "Character Count": char_count
+        }
+
+    # Gradio UI
+    with gr.Blocks(title="Whisper Agentic Transcriber") as demo:
+        gr.Markdown("## 🤖 Whisper + Agentic Analysis")
+        audio = gr.Audio(sources=["microphone", "upload"], type="filepath", label="Record Audio")
+        btn = gr.Button("Transcribe and Analyze")
+        transcript = gr.Textbox(label="📝 Transcript", lines=4)
+        word_count = gr.Number(label="📏 Word Count")
+        char_count = gr.Number(label="🔡 Character Count")
+
+        btn.click(fn=transcribe_and_analyze, inputs=audio, outputs=[transcript, word_count, char_count])
+
+    demo.launch(server_name="0.0.0.0", server_port=7860, share=True)