How-to: Detecting secrets in video content

GitLab has open-sourced a custom solution for detecting secrets in video content, specifically designed to identify API keys and sensitive tokens within videos on their GitLab Unfiltered YouTube channel. The approach involves splitting videos into frames, using optical character recognition (OCR) to extract text, and matching the text against known secret patterns. Initially, they experimented with FFmpeg and Tesseract for frame splitting and OCR but ultimately chose Google's Video Intelligence API for its ease of use despite its limitations in control. The solution utilizes approximate regular expression matching to account for OCR inaccuracies, allowing for small deviations in detected text. GitLab's AI assistant, Duo Chat, aided in developing a Node.js addon to integrate TRE, a C library for approximate regular expression matching, into the scanner. This open-source project invites feedback and aims to enhance video content security by identifying and mitigating potential secret leaks.

Use this post, company, and trend context to find content marketing opportunities, perform competitive analysis, or address product feature gaps via the Plushcap MCP server or the Plushcap API.