social-analyzer

Social Analyzer is an open-source OSINT tool for analyzing and locating a person's profile across more than 1000 social media sites and websites. It is distributed as an API, a command-line interface, and a local web application, and it is intended to be run locally rather than as a hosted service. The project uses multiple detection and analysis modules. Detection relies on a rating mechanism combining different techniques, producing a score from 0 to 100 (No, Maybe, Yes) to limit false positives. Analysis and publicly extracted information are positioned as aids for investigating profiles tied to suspicious or malicious activity such as cyberbullying, cyber grooming, cyberstalking, and misinformation. The README notes the project is used by some law enforcement agencies in resource-limited countries, and that their detection database differs from the public one. Features listed include string and name analysis with permutations and combinations; profile discovery via HTTPS libraries and Webdriver; multi-profile search for correlation using comma-separated names; multilayer detection (OCR, normal, advanced, special); visualized profile information through Ixora with metadata and pattern extraction; a force-directed metadata graph; search by top ranking or country using Alexa ranking; search by website type such as adult or music; profile statistics and static info; cross-metadata stats; search engine lookup via optional Google API; custom queries via Google and DuckDuckGo APIs; profile screenshots, titles, info, and site descriptions; name origin, similarity, and common-word analysis by language; limited age estimation; configurable user-agent, proxy, timeout, and implicit wait; Python and NodeJS CLIs; screenshots of detected profiles requiring a recent Chrome; a grid option for faster checking limited to docker-compose; log dumping to folder or terminal; adjustable worker counts (default 15); re-checking of failed profiles; filtering by good, maybe, or bad; JSON export; and a simplified web interface and CLI. Special detections are listed for Facebook (phone number, name, or profile name), Gmail addresses, and Google accounts. Installation is documented for Linux as a Node web app, Node CLI, Python package, or Python script, with additional methods for Windows, macOS, and Raspberry Pi referenced in the wiki. The web app runs at localhost on port 9005. Command-line options cover username, website selection, analysis mode (fast, slow, special), output format, displayed fields, method, filters, countries, type, top-N selection, extraction, metadata, trimming, headers, logs directory, timeout, and silent mode. The tool can also be imported as a Python object. Resources include DuckDuckGo API, Google API, NodeJS, bootstrap, selectize, jQuery, Wikipedia, font-awesome, selenium-webdriver, and tesseract.js. The README includes a disclaimer that the project should be treated as a security project, that websites can request exclusion, that it lacks access control and is meant for local use, and that issues with private modules should be raised directly rather than on GitHub.