Show HN: Keyfence is a local proxy that stops secrets from reaching LLM APIs (github.com)

Keyfence, a newly launched local proxy tool, aims to enhance security by preventing sensitive information, such as API keys and passwords, from being transmitted to large language model (LLM) APIs. This tool inspects every request before it leaves the user's machine, allowing users to block, redact, or replace secrets with placeholders. Keyfence is compatible with a variety of platforms including Claude Code, Cursor, and Codex, making it a versatile solution for developers working with LLMs through HTTP communication. The significance of Keyfence lies in its ability to safeguard proprietary information while utilizing AI services, thereby addressing growing concerns around data leaks in AI applications. With features allowing users to register secrets and test with dummy data, this tool empowers developers to maintain control over sensitive information, enhancing privacy and security in AI workflows. Its reliance on established tools like `mitmproxy` for HTTPS inspection underscores the technical rigor while providing a straightforward setup. As organizations increasingly adopt AI technologies, solutions like Keyfence could become essential in fostering a security-conscious environment within the AI/ML community.