scan-for-webcams  by JettChenT

CLI tool for scanning webcams on the internet

created 5 years ago
259 stars

Top 98.4% on sourcepulse

GitHubView on GitHub
Project Summary

This project provides a tool for scanning the internet for publicly accessible webcams, targeting security researchers and hobbyists interested in network reconnaissance. It simplifies the discovery and viewing of live streams from various webcam types, offering both command-line and experimental GUI interfaces.

How It Works

The tool leverages Shodan for initial discovery of internet-connected devices, specifically targeting webcam protocols like MJPG, webcamXP, yawCam, hipcam, and RTSP. It then enumerates and attempts to display streams, with an option to use a Places365 model for on-device location classification of footage, enhancing the descriptive capabilities beyond simple tags. Experimental support for Vision-Language Models (VLMs) like LLaVA is also included for natural language descriptions.

Quick Start & Requirements

  • Install via pip install -r requirements.txt after cloning the repository.
  • Requires API keys for Shodan, Clarifai (or local Places365 model), and geo.ipify.org.
  • Setup involves running python sfw setup to input API keys.
  • Experimental VLM support requires llama-cpp-python and huggingface_hub.
  • Official Docs: https://github.com/JettChenT/scan-for-webcams

Highlighted Details

  • Supports multiple webcam stream types (MJPG, webcamXP, yawCam, hipcam, RTSP).
  • Optional GUI for visual display of scanned streams.
  • Experimental on-device footage classification using Places365.
  • Experimental Vision-Language Model integration for natural language descriptions.

Maintenance & Community

  • The project maintains a Discord channel for community interaction.
  • The README notes that the PyPI package is deprecated in favor of direct repository installation for better developer experience.

Licensing & Compatibility

  • The repository does not explicitly state a license in the provided README.

Limitations & Caveats

  • The Vision-Language Model integration is experimental and requires disabling parallel processing.
  • RTSP stream enumeration is noted as potentially dangerous.
  • The project relies on external API services (Shodan, Clarifai, geo.ipify.org) which may have usage limits or costs.
Health Check
Last commit

1 year ago

Responsiveness

1+ week

Pull Requests (30d)
0
Issues (30d)
0
Star History
6 stars in the last 90 days

Explore Similar Projects

Feedback? Help us improve.