Show HN: Aitori, see and govern the AI traffic leaving your machine (github.com)

Aitori is a new tool designed to intercept and govern AI traffic leaving a device, acting as a gatekeeper for model and MCP calls. It functions as an agent running locally, logging requests through a specified gateway while leaving other traffic unaffected. Users install aitori easily on macOS or Linux, and with no additional configuration, it can decrypt and monitor calls made to prominent AI services like Claude and ChatGPT. This functionality enables users to inspect, log, and enforce policies on their AI interactions without breaking application performance. The significance of aitori in the AI/ML community lies in its potential to enhance transparency and governance around AI usage on personal devices. By allowing users to have oversight and control over how their requests are routed and processed, aitori addresses a critical gap in existing tools, which often lack the ability to manage AI traffic effectively. Its reliance on a per-device Certificate Authority (CA) ensures that sensitive keys remain secure, while the tool also offers extensibility for developers to tailor configurations to their needs. Aitori represents an important step towards fostering responsible AI usage in everyday applications.