Skip to content

DEPENDENCIES.md

Latest commit

 

History

History
131 lines (98 loc) · 4.66 KB

DEPENDENCIES.md

File metadata and controls

131 lines (98 loc) · 4.66 KB

Dependencies

All dependencies are installed automatically by scripts/install.sh. This document lists them for reference.

External Services (Optional)

These are separate projects that can be self-hosted or used as public services:

Service Purpose Options
ntfy Push notifications to mobile Use public ntfy.sh or self-host
bore TCP tunneling for remote access Self-host server + local client

ntfy Setup

For push notifications, you have two options:

Option A: Public ntfy.sh (easiest)

  • No setup required
  • Set NtfyUrl: "https://ntfy.sh" in config
  • Pick a unique topic name

Option B: Self-hosted ntfy

# Install ntfy server
curl -sSL https://install.ntfy.sh | sh

# Or via Docker
docker run -p 8085:80 binwiederhier/ntfy serve

bore Setup (for remote access)

If you want to access the viewer from outside your network:

# Install bore client (on arlo-base)
cargo install bore-cli

# On your public server, run bore server
bore server --min-port 8080 --max-port 8090

# bore tunnels are configured in install.conf

System Packages (apt)

Package Purpose
hostapd WiFi access point daemon
dnsmasq DHCP and DNS server for camera network
netfilter-persistent Persistent iptables firewall rules
iptables-persistent Saves/restores firewall on boot
python3-venv Python virtual environment support
python3-full Complete Python 3 installation
ffmpeg Video processing (thumbnails, format conversion)
nodejs Runtime for arlo-viewer web server
npm Node.js package manager
gstreamer1.0-tools GStreamer CLI tools (gst-launch-1.0)
gstreamer1.0-plugins-base Core GStreamer plugins
gstreamer1.0-plugins-good Quality GStreamer plugins (RTSP, HLS)
gstreamer1.0-plugins-bad Additional GStreamer plugins
python3-gst-1.0 Python bindings for GStreamer

Optional

Package Purpose
bore TCP tunnel for remote access (install via cargo)

Python Packages (pip)

Located in src/arlo-cam-api/requirements.txt:

Package Version Purpose
Flask 1.1.2 Web framework for REST API
PyYAML 5.3.1 YAML config file parsing
pyaml 20.4.0 YAML utilities
requests 2.25.0 HTTP client for webhooks/ntfy
webhooks 0.4.2 Webhook dispatch
python-vlc 3.0.11115 VLC bindings (legacy, may not be used)
Jinja2 2.11.2 Flask templating
Werkzeug 1.0.1 Flask WSGI utilities
click 7.1.2 CLI framework (Flask dependency)
itsdangerous 1.1.0 Flask session signing
MarkupSafe 1.1.1 Jinja2 dependency
certifi 2020.11.8 SSL certificates
chardet 3.0.4 Character encoding detection
idna 2.10 Internationalized domain names
urllib3 1.26.2 HTTP library
wrapt 1.12.1 Decorator utilities
standardjson 0.3.1 JSON utilities
cached-property 1.5.2 Cached property decorator

Node.js Packages (npm)

Located in src/arlo-viewer/package.json:

Package Version Purpose
express ^4.18.2 Web server framework
js-yaml ^4.1.1 YAML parsing for config

Why GStreamer Instead of FFmpeg?

Both are installed, but GStreamer is used for live streaming because:

  • FFmpeg sends RTCP keep-alive packets every 10 seconds (hardcoded)
  • Arlo cameras require RTCP every 5 seconds or they stop streaming
  • GStreamer correctly sends RTCP at 5-second intervals

FFmpeg is still used for thumbnail generation from recorded videos.

Version Notes

The Python dependencies use older versions for compatibility with the original arlo-cam-api fork. Consider updating for security patches, but test thoroughly as Flask 1.x → 2.x has breaking changes.

Credits / Based On

Project Author Purpose
arlo-cam-api Meatballs1 Original Arlo camera protocol implementation
ntfy binwiederhier Push notification server
bore ekzhang TCP tunnel for remote access

The arlo-cam-api project reverse-engineered the Arlo camera protocol and provided the foundation for this base station replacement. This project extends it with:

  • GStreamer-based live streaming (fixes RTCP timing issues)
  • Web-based recording viewer
  • ntfy push notifications with thumbnails
  • Improved WiFi hardware compatibility documentation