eye_of_web
EyeOfWeb is an open-source OSINT platform for web-based facial intelligence and security analysis. Built with Python, Flask, PostgreSQL, Milvus, and InsightFace, it serves as an alternative to commercial face search engines. Key features include facial recognition using the AntelopeV2 model, a web crawler for collecting images from open sources, vector similarity search via Milvus, and a REST API for integration. The platform is designed for academic research, education, and legal security simulations. Users can deploy it quickly via Docker, configure crawl targets and sources, and manage collected data through a structured pipeline. Note: Docker images ship with Torch CPU by default; GPU usage requires modifying the Dockerfile. Documentation is available in Turkish, English, Russian, and Chinese. Licensed under MIT. The developers emphasize that the tool must only be used for lawful purposes and that end-users bear full responsibility for compliance with privacy laws such as GDPR and KVKK.