Upload 3 files

README.md

@@ -0,0 +1,43 @@
+---
+title: EchoScribe
+emoji: 🎥
+colorFrom: indigo
+colorTo: pink
+sdk: gradio
+sdk_version: "3.50.2"
+app_file: app.py
+pinned: false
+---
+
+# 🎥 EchoScribe: Smart Video Transcriber
+
+**EchoScribe** is a powerful AI tool that turns your videos into clean, punctuated transcripts, live subtitles, and insightful summaries — all in one click.
+
+## 🚀 Features
+
+- 🎬 Upload any video (MP4)
+- 🧾 Get raw and punctuated transcripts using `openai/whisper-large` and `oliverguhr/fullstop-punctuation-multilang-large`
+- 📝 Generate clean summaries using `facebook/bart-large-cnn`
+- ⬇️ Download transcript, summary, and subtitle (.srt) files
+- 🌈 Clean, responsive interface powered by Gradio
+- 🖤 Built with ❤️ by Snigdha’s AI Lab
+
+## 🔧 Technologies Used
+
+- Hugging Face Transformers
+- Gradio
+- Whisper ASR
+- MoviePy for audio extraction
+- Python, Torch
+
+## 🛠️ Usage
+
+1. Upload a short video clip
+2. Wait for the transcript and summary to be generated
+3. View and download the results instantly
+
+> Ideal for note-taking, content summarization, interviews, YouTube creators, and accessibility projects.
+
+---
+
+Feel free to fork, contribute, and remix!

app.py

@@ -0,0 +1,96 @@
+import gradio as gr
+from transformers import pipeline
+import tempfile
+import torch
+import os
+from moviepy.editor import VideoFileClip
+import srt
+import datetime
+
+# Load Hugging Face models
+device = 0 if torch.cuda.is_available() else -1
+whisper = pipeline("automatic-speech-recognition", model="openai/whisper-large", device=device)
+punctuate = pipeline("text2text-generation", model="oliverguhr/fullstop-punctuation-multilang-large")
+summarizer = pipeline("summarization", model="facebook/bart-large-cnn")
+
+# Utility: Extract audio
+def extract_audio(video_path):
+    video = VideoFileClip(video_path)
+    audio_path = tempfile.mktemp(suffix=".wav")
+    video.audio.write_audiofile(audio_path, verbose=False, logger=None)
+    return audio_path
+
+# Utility: Create .srt subtitles
+def generate_srt(transcript_text):
+    lines = transcript_text.strip().split(". ")
+    subs = []
+    for i, line in enumerate(lines):
+        start = datetime.timedelta(seconds=i*2)
+        end = datetime.timedelta(seconds=(i+1)*2)
+        subs.append(srt.Subtitle(index=i+1, start=start, end=end, content=line.strip()))
+    srt_data = srt.compose(subs)
+    srt_path = tempfile.mktemp(suffix=".srt")
+    with open(srt_path, "w") as f:
+        f.write(srt_data)
+    return srt_path
+
+# Full pipeline
+def transcribe_pipeline(video_file):
+    with tempfile.NamedTemporaryFile(delete=False, suffix=".mp4") as tmp:
+        tmp.write(video_file.read())
+        video_path = tmp.name
+
+    audio_path = extract_audio(video_path)
+    result = whisper(audio_path)
+    raw_text = result["text"]
+
+    punctuated = punctuate(raw_text)[0]["generated_text"]
+    summary = summarizer(punctuated, max_length=130, min_length=30, do_sample=False)[0]['summary_text']
+    srt_path = generate_srt(punctuated)
+
+    # Save raw text, punctuated text, and summary
+    raw_txt_path = tempfile.mktemp(suffix=".txt")
+    punct_txt_path = tempfile.mktemp(suffix=".txt")
+    summary_txt_path = tempfile.mktemp(suffix=".txt")
+
+    with open(raw_txt_path, "w") as f:
+        f.write(raw_text)
+    with open(punct_txt_path, "w") as f:
+        f.write(punctuated)
+    with open(summary_txt_path, "w") as f:
+        f.write(summary)
+
+    return raw_text, punctuated, summary, punct_txt_path, summary_txt_path, srt_path
+
+# Gradio UI
+with gr.Blocks(theme=gr.themes.Soft()) as iface:
+    gr.Markdown("# 🎥 EchoScribe: Smart Video Transcriber")
+    gr.Markdown("Upload a video to extract transcript with punctuation and summary using Hugging Face models. Powered by Whisper, BART, and punctuation restoration.")
+
+    with gr.Row():
+        video_input = gr.Video(label="🎬 Upload your video")
+
+    with gr.Row():
+        raw_output = gr.Textbox(label="🧾 Raw Transcript (Whisper)", lines=6)
+        punct_output = gr.Textbox(label="📄 Punctuated Transcript", lines=6)
+
+    summary_output = gr.Textbox(label="📝 Summary", lines=4)
+
+    with gr.Row():
+        download_transcript = gr.File(label="⬇️ Download Transcript (.txt)")
+        download_summary = gr.File(label="⬇️ Download Summary (.txt)")
+        download_srt = gr.File(label="⬇️ Download Subtitles (.srt)")
+
+    submit_btn = gr.Button("🚀 Transcribe & Summarize")
+
+    def run_all(video):
+        return transcribe_pipeline(video)
+
+    submit_btn.click(fn=run_all, inputs=video_input,
+                     outputs=[raw_output, punct_output, summary_output,
+                              download_transcript, download_summary, download_srt])
+
+    gr.Markdown("---")
+    gr.Markdown("🔧 Built with ❤️ by **Snigdha’s AI Lab**")
+
+iface.launch()

requirements.txt

@@ -0,0 +1,7 @@
+gradio
+transformers
+torch
+moviepy
+pydub
+ffmpeg-python
+srt