Snortstamper Web UI - Setup Guide

Snortstamper uses a locally-installed AI model to generate timestamps for any length of youtube transcripts.

Installation Steps

1. Prerequisites

Ensure you have:

• Python 3.8+

• Ollama installed

• Ollama running with Mistral model: ollama pull mistral

• Get the transcript

  • For the correct timestamp format, you need the VidIQ plugin. https://app.vidiq.com/

• Go to any youtube video

  • Go to the description -> transcript
  • Now you will see the VidIQ plugin button to copy the transcript
  • Press the dropdown button and copy the transcript with timestamps
  • Paste the whole transcript into the transcript.txt file

Running the Application

Easiest Way: One-Click Startup (Recommended)

Windows:

1. Double-click Run Snortstamper.ink
2. Wait for browser to prompt
3. Open http://localhost:5000

macOS/Linux:

1. Open terminal in the project folder
2. Run: chmod +x startup.sh && ./startup.sh
3. Wait for Flask to start
4. Open http://localhost:5000

The startup script will:

• Create virtual environment (if needed)
• Activate it
• Check for Ollama and Mistral model
• Start Ollama server
• Start Flask server
• Open the UI

---

Manual Setup (if you prefer)

Step 1: Activate Virtual Environment

Windows:

venv\Scripts\activate

macOS/Linux:

source venv/bin/activate

Step 2: Start Ollama (in one terminal)

ollama serve

Step 3: Start Flask Server (in another terminal)

python app.py

Step 4: Open in Browser Go to: http://localhost:5000

Usage

1. Click on the upload area to select your transcript file (.txt)
2. The file should contain timestamps in [M:SS] or [H:MM:SS] format
3. Click "Generate Timestamps"
4. Wait for processing (will print progress in terminal)
5. View the generated chapters in the preview
6. Click "Download Chapters" to save as youtube_chapters.txt

Command-Line Usage (Still Supported)

If you want to use the script from the command line without the web UI:

from snortstamper_core import ChapterGenerator
import json

with open('transcript.txt', 'r', encoding='utf-8') as f:
    transcript = f.read()

generator = ChapterGenerator(model="mistral")
chapters = generator.generate_chapters(transcript)
formatted = generator.format_chapters(chapters)

print(formatted)

# Optionally save
with open('youtube_chapters.txt', 'w', encoding='utf-8') as f:
    f.write(formatted)

with open('chapters.json', 'w', encoding='utf-8') as f:
    json.dump(chapters, f, indent=2, ensure_ascii=False)

Troubleshooting

Error: "Connection refused" on port 5000

• Flask server isn't running. Run python app.py

Error: "Could not connect to Ollama"

• Ollama isn't running. Run ollama serve in another terminal
• Make sure Mistral is pulled: ollama pull mistral

Error: "No transcript file provided"

• Make sure you selected a .txt file before clicking Generate

Long processing time

• This is normal! LLM processing takes time. Check the terminal for progress logs.

API Endpoint

POST /api/generate-timestamps

Request:
- Form data with file field named "transcript" (multipart/form-data)

Response:
{
  "timestamps": "[0:00] Chapter Title\n[1:30] Another Chapter\n..."
}

Error Response:
{
  "error": "Error message"
}