ScrybeSync

ScrybeSync is a personal transcription and note-taking web application. It converts audio/video recordings into transcripts, generates summaries, and lets users organize the results as notes.

Product Focus

ScrybeSync is a practical personal productivity tool, not a commercial SaaS application. It is designed for straightforward personal workflows:

• Uploading audio and video recordings
• Transcribing speech using OpenAI Whisper
• Generating summaries from completed transcripts
• Creating and organizing personal markdown notes

---

Features

Implemented

• User Authentication: Simple user registration, login, logout, HttpOnly refresh cookie rotation, and an authenticated /api/users/me/ endpoint to restore user state.
• Media Upload: File upload support for common audio formats (.mp3, .wav, .m4a, .flac, .ogg) and video formats (.mp4, .avi, .mov, .mkv, .webm).
• Asynchronous Transcription: Background transcription worker via Celery and OpenAI Whisper.
• Retry Mechanism: Dedicated retry endpoint strictly for failed transcriptions.
• Transcript Summarization: Asynchronous structured summary generation for completed transcripts using DeepSeek V4 Flash via the official OpenAI Python SDK.
• Notes Management: Create, view, update, and delete personal notes.
• User-Isolated Categories: Categories are scoped per user (e.g. two users can both create a "School" category, but a single user cannot create duplicates). Server-side validation prevents cross-user category assignment.
• Practical Workspace: Minimal dashboard featuring recent transcriptions, recent notes, and real item counts.
• Theme Support: Light and dark mode with persistent user preference.

Planned

• URL Import: Direct media importing from YouTube and other video/audio platforms using yt-dlp.
• Interactive Transcript Viewer: Timed segment viewer with inline note-taking and section clipping.
• Bi-directional Note Linking: Interconnected personal wiki-style note references.

---

Tech Stack

Backend

• Framework: Django 5.2 & Django REST Framework
• Database: PostgreSQL (with SQLite support for local development/testing)
• Task Queue: Celery & Redis
• Transcription Engine: OpenAI Whisper & FFmpeg
• LLM Integration: DeepSeek V4 Flash (via OpenAI Python SDK)

Frontend

• Framework: React 19 & Vite 6
• Routing: React Router 7
• Styling: Tailwind CSS 4
• State Management: Zustand
• HTTP Client: Axios

---

API Endpoints

Authentication & Users

Method	Endpoint	Description
POST	/api/users/register/	Register a new user
POST	/api/users/login/	Authenticate with credentials; returns access token and sets HttpOnly refresh cookie
POST	/api/users/token/refresh/	Refresh access token using HttpOnly cookie
POST	/api/users/logout/	Invalidate refresh token and clear session
GET	/api/users/me/	Retrieve currently authenticated user profile

Transcripts

Method	Endpoint	Description
GET	/api/transcriber/transcripts/	List current user's transcripts
POST	/api/transcriber/transcripts/	Upload audio/video file and queue transcription
GET	/api/transcriber/transcripts/{id}/	Retrieve transcript details and text
POST	/api/transcriber/transcripts/{id}/retry_transcription/	Retry a failed transcription
GET	/api/transcriber/transcripts/status_summary/	Get transcript count summary by status
DELETE	/api/transcriber/transcripts/{id}/	Delete a transcript

Summaries

Method	Endpoint	Description
GET	/api/summarizer/summaries/	List current user's summaries
POST	/api/summarizer/summaries/	Generate summary for a completed transcript
GET	/api/summarizer/summaries/{id}/	Retrieve summary details
POST	/api/summarizer/summaries/{id}/regenerate/	Regenerate summary content
GET	/api/summarizer/summaries/status_summary/	Get summary count summary by status
DELETE	/api/summarizer/summaries/{id}/	Delete a summary

Notes & Categories

Method	Endpoint	Description
GET	/api/notes/notes/	List current user's notes
POST	/api/notes/notes/	Create a note (server validates category ownership)
GET	/api/notes/notes/{id}/	Retrieve note details
PATCH	/api/notes/notes/{id}/	Update note content or category
DELETE	/api/notes/notes/{id}/	Delete a note
GET	/api/notes/categories/	List current user's categories
POST	/api/notes/categories/	Create a category (scoped uniquely per user)
DELETE	/api/notes/categories/{id}/	Delete a category

---

Configuration & Environment Variables

Backend (backend/.env)

SECRET_KEY=your_django_secret_key
DEBUG=True
ALLOWED_HOSTS=localhost,127.0.0.1

# Database (defaults to PostgreSQL, or set USE_SQLITE=True for local dev/testing)
USE_SQLITE=False
DB_ENGINE=django.db.backends.postgresql
DB_NAME=scrybesync_db
DB_USER=postgres
DB_PASSWORD=postgres
DB_HOST=localhost
DB_PORT=5432

# Celery & Redis
CELERY_BROKER_URL=redis://localhost:6379/0
CELERY_RESULT_BACKEND=redis://localhost:6379/0

# AI & Media
DEEPSEEK_API_KEY=your_deepseek_api_key
DEEPSEEK_BASE_URL=https://api.deepseek.com
SUMMARY_MODEL=deepseek-v4-flash
WHISPER_MODEL=base
WHISPER_DEVICE=cpu
MAX_UPLOAD_SIZE=100MB

Frontend (frontend/.env)

VITE_API_BASE_URL=http://localhost:8000/api

---

Running Locally

Backend

cd backend
python -m venv .venv
source .venv/bin/activate  # Or .venv\Scripts\activate on Windows
pip install -r requirements.txt
python manage.py migrate
python manage.py runserver

To run Celery worker:

# Linux / macOS
celery -A backend worker -l info

# Windows (requires solo pool due to lack of fork support)
celery -A backend worker -l info --pool=solo

To run backend tests:

python manage.py test

Frontend

cd frontend
npm install
npm run dev

To run frontend lint and build:

npm run lint
npm run build

---

License

MIT