ytgrid 3.1.0

YTGrid is a powerful, scalable, and flexible YT automation tool that enables looped playback, remote control, and real-time tracking using a hybrid CLI + API architecture. It integrates FastAPI for REST API control, Selenium for browser automation, and Python multiprocessing/Celery for concurrent tasks.

🎥 YTGrid — Hybrid CLI + API for Scalable YT Automation

YTGrid is a powerful, scalable, and flexible YouTube automation tool that enables looped playback, remote control, and real-time tracking through a hybrid CLI + API architecture.

It combines FastAPI for REST control, Selenium for browser automation, Python multiprocessing or Celery for concurrent execution, and a SQLite-backed persistence layer for reusable execution profiles.

---

✨ Features

Core

• Hybrid Interface – manage automation via CLI, REST API, or web dashboard.
• Scalable Execution – run multiple browser instances in parallel via multiprocessing or Celery.
• Configurable Automation – playback speed, loop count, task type.
• Real-time Updates – live session status via Server-Sent Events.

v3.1 additions

• 🔐 API key authentication – every endpoint protected by X-API-Key header (or api_key query param for browser EventSource).
• 🗂️ Execution profiles – save named playlists of URLs + parameters in SQLite; rerun on demand from the CLI or API.
• 🎛️ Web dashboard – live session table and task launcher at /static/index.html; advanced analytics remain on the roadmap.
• 🌐 Optional proxy rotation – file/env proxy sources, weighted selection, and background health checks; dashboard proxy panels remain on the roadmap.
• 🧹 Resource optimizer – background tmp-dir cleaner, orphan-only zombie reaper (ppid==1 Chrome processes), system-load throttling of new sessions.
• 📺 Playlist & channel automation – new task_type=playlist / channel; URLs resolved via yt-dlp.
• 🐳 Split Docker images – Dockerfile.api (FastAPI only) and Dockerfile.worker (Chrome + Celery worker) for microservice deployment.
• 🛡️ Hardened defaults – input validation, process-group cleanup, isolated browser profiles (0o700), WAL journaling, lifespan-managed startup/shutdown.

---

📦 Installation

1️⃣ PyPI

pip install ytgrid

2️⃣ Docker

docker pull sandy1sp/ytgrid:latest
docker run -p 8000:8000 -e YTGRID_API_KEY=$(openssl rand -hex 32) sandy1sp/ytgrid:latest

3️⃣ From source (development)

git clone https://github.com/sandy-sp/ytgrid.git
cd ytgrid
poetry install
cp .env.example .env
# generate an API key
python -c "import secrets; print(secrets.token_urlsafe(32))"
# paste it into .env as YTGRID_API_KEY=...

Requirements

• Python 3.10+
• Google Chrome – required for the PyPI install; bundled in the Docker image.
• ChromeDriver – managed automatically by webdriver-manager.
• Redis – only needed if Celery is enabled.

---

🔐 Authentication

When YTGRID_API_KEY is set, every endpoint (except /, /health, and static assets) requires the key:

curl -H "X-API-Key: $YTGRID_API_KEY" http://127.0.0.1:8000/tasks/

The CLI reads the key from the same .env and forwards it automatically. Browser dashboard EventSource connections may pass the key as ?api_key=<key> since browsers cannot set custom headers on EventSource.

When YTGRID_API_KEY is unset, auth is disabled (useful for local development).

---

🚀 CLI Usage

Start a session

ytgrid start --url "https://www.youtube.com/watch?v=UXFBUZEpnrc" --speed 1.5 --loops 3

--session-id is now optional — a short id is auto-generated when omitted.

Status

ytgrid status

Stop a session

ytgrid stop --session-id abc12345

Batch (CSV)

ytgrid batch tasks.csv --delimiter ","

CSV header: session_id, url, speed, loops, task_type.

Profiles (new in v3.1)

ytgrid profile create "morning-mix" --description "Daily wake-up playlist"
ytgrid profile add   "morning-mix" --url "https://www.youtube.com/watch?v=UXFBUZEpnrc" --loops 2
ytgrid profile add   "morning-mix" --url "https://www.youtube.com/watch?v=OaOK76hiW8I" --speed 1.5
ytgrid profile list
ytgrid profile run   "morning-mix"

Dashboard server

ytgrid server
# → http://127.0.0.1:8000/static/index.html

Toggle Celery

ytgrid toggle-celery

---

🖥️ REST API

All examples assume YTGRID_API_KEY is set; substitute your own key.

Start a task

curl -X POST http://127.0.0.1:8000/tasks/ \
     -H "X-API-Key: $YTGRID_API_KEY" \
     -H "Content-Type: application/json" \
     -d '{"url": "https://www.youtube.com/watch?v=OaOK76hiW8I", "speed": 1.5, "loop_count": 3}'

List active tasks

curl -H "X-API-Key: $YTGRID_API_KEY" http://127.0.0.1:8000/tasks/

Stop a task

curl -X POST http://127.0.0.1:8000/tasks/stop \
     -H "X-API-Key: $YTGRID_API_KEY" \
     -H "Content-Type: application/json" \
     -d '{"session_id": "abc12345"}'

Profiles

# create
curl -X POST http://127.0.0.1:8000/profiles/ \
     -H "X-API-Key: $YTGRID_API_KEY" -H "Content-Type: application/json" \
     -d '{"name": "morning-mix", "description": "wake up"}'

# add entry
curl -X POST http://127.0.0.1:8000/profiles/1/entries \
     -H "X-API-Key: $YTGRID_API_KEY" -H "Content-Type: application/json" \
     -d '{"video_url": "https://www.youtube.com/watch?v=OaOK76hiW8I", "speed": 1.0, "loop_count": 2}'

# run all entries
curl -X POST http://127.0.0.1:8000/profiles/morning-mix/run \
     -H "X-API-Key: $YTGRID_API_KEY"

Streaming endpoints

Endpoint	Purpose
GET /tasks/stream	SSE — active session updates every 5 s (auth required)
GET /dashboard/stream	SSE — dashboard payload every 2 s (auth required; supports ?api_key=)

Both cap at 10 concurrent connections and 1 h max duration.

---

🎛️ Web Dashboard

Open http://127.0.0.1:8000/static/index.html after starting the server. The dashboard shows live sessions, system health, and a "Launch Task" form. On first load it prompts for the API key (stored in localStorage).

For the full shipped-app summary, current limitations, and release evidence, see docs/APP_OVERVIEW.md.

---

🌐 Proxy Rotation (optional)

Enable in .env:

YTGRID_PROXY_ENABLED=True
YTGRID_PROXY_SOURCE=file
YTGRID_PROXY_FILE=./proxies.txt

proxies.txt format (one per line):

host:port
host:port:user:pass

The pool selects proxies by a weighted score (latency × reliability × idle-time), runs a background health checker, and reports per-session success/failure.

---

🧹 Resource Optimizer

Runs three background threads when the API is up:

Job	Interval	Action
TmpCleaner	5 min	Deletes /tmp/ytgrid_* dirs older than YTGRID_TMP_MAX_AGE
ZombieReaper	1 min	SIGTERMs orphaned Chrome processes (ppid==1, current UID, >10 min old)
SystemMonitor	30 s	Samples CPU/mem/disk; throttles new sessions when load is high

Disable with YTGRID_OPTIMIZER_ENABLED=False.

---

🐳 Docker

Single container

docker build -t ytgrid .
docker run -p 8000:8000 -e YTGRID_API_KEY=$(openssl rand -hex 32) ytgrid

Split API + worker (recommended)

docker-compose up --build

This brings up:

• ytgrid_api — FastAPI gateway (Dockerfile.api)
• ytgrid_worker — Celery worker with Chrome (Dockerfile.worker)
• redis — broker / result backend

Docker Compose requires an API key because it publishes the API port:

export YTGRID_API_KEY=$(openssl rand -hex 32)
docker-compose up --build

---

🛠️ Configuration (.env)

See .env.example for the full template.

Variable	Default	Purpose
YTGRID_API_KEY	(empty = auth disabled)	API authentication key
YTGRID_HEADLESS_MODE	True	Chrome headless toggle
YTGRID_DEFAULT_SPEED / YTGRID_DEFAULT_LOOP_COUNT	1.0 / 1	CLI defaults
YTGRID_MAX_SESSIONS	5	Concurrency cap
YTGRID_BROWSER_TIMEOUT	20	Selenium implicit wait (s)
YTGRID_DB_PATH	ytgrid/ytgrid.db	SQLite path for profiles + history
YTGRID_USE_CELERY	False	Use Celery instead of multiprocessing
CELERY_BROKER_URL / CELERY_RESULT_BACKEND	redis://localhost:6379/0	Celery transport
YTGRID_OPTIMIZER_ENABLED	True	Enable background optimizer
YTGRID_TMP_MAX_AGE	1800	Stale tmp-dir age threshold (s)
YTGRID_PROXY_ENABLED	False	Route Chrome through rotating proxies
YTGRID_PROXY_SOURCE	file	file / env / api
YTGRID_PROXY_FILE	./proxies.txt	Proxy list path
YTGRID_PROXY_COOLDOWN_SECONDS	300	Per-proxy cooldown after selection
YTGRID_PROXY_HEALTH_CHECK_INTERVAL	60	Health-check loop interval (s)
YTGRID_PROXY_MAX_FAILURE_RATE	0.3	Health threshold

---

🧪 Testing

poetry install --with dev
poetry run pytest

For step-by-step manual end-to-end checks see docs/MANUAL_TESTING.md.

Architectural docs:

• docs/VULNERABILITY_FIXES.md
• docs/FEATURE_SPECIFICATIONS.md
• docs/IMPLEMENTATION_PLAN.md
• docs/ROADMAP_SPECIFICATIONS.md

---

🔮 Roadmap

• Execution-history UI for GET /profiles/{id}/history
• Proxy stats endpoint and dashboard panel
• Profile import/export (JSON)
• Kubernetes manifests for the split-microservice deployment
• Optional PostgreSQL backend for multi-node profile sharing

---

📜 License

MIT

---

🌍 Contributing

1. Fork the repo
2. git checkout -b feature/my-feature
3. Add tests for your change (pytest must pass)
4. Open a PR

---

📖 Resources

• PyPI
• Docker Hub
• GitHub