Octopii

Octopii is an open source scanner that finds personally identifiable information in images. It was released by the security company RedHunt Labs in 2022 as a free alternative to commercial tools, after the team found that Amazon Macie only covers S3 and ManageEngine DataSecurity Plus is paid. It combines two techniques. A MobileNet convolutional neural network built with Keras classifies each image by document type, covering items such as government IDs, passports, debit cards, driver's licenses, photos and signatures. Tesseract OCR then reads text from the same image and checks it against search strings that identify each type, for example a word found on a bank passbook and not on a license. The search strings can be extended with regular expressions to improve accuracy. It is a Python command line tool. A user installs the dependencies from a requirements file and runs it against a target, which can be a local file system path, an Amazon S3 bucket or an open directory listing reached by URL. The intended users are security and compliance teams that need to find exposed KYC documents and other sensitive images in storage they own, including misconfigured buckets. The listing points at the vendor's announcement article for the tool, which carries the description and installation steps, and the code is published on GitHub. The tool is free and open source, and no price is shown for it. RedHunt Labs separately sells an attack surface management platform with demo and free trial options, but the site shows no prices for that either.