Show HN: AgentLens – Open-source observability and audit trail for AI agents (github.com)

AgentLens has been announced as an open-source tool designed to enhance the observability of AI agents by capturing every interaction, decision, and error encountered during operation. This innovative “flight recorder” solution enables users to monitor AI agents in real time through a user-friendly dashboard and queryable API. With its zero-code setup that integrates seamlessly into existing environments, AgentLens simplifies the process of gathering detailed analytics, including cost tracking and generating alerts for various performance metrics, such as error rates and latency anomalies. The significance of AgentLens lies in its ability to strengthen the transparency and accountability of AI systems, which are often characterized by complex decision-making processes. By providing a tamper-evident audit trail with SHA-256 hash chains and session-based storage, it ensures that all recorded events can be verified, enhancing trust among developers and stakeholders. This tool operates with a self-hosted SQLite backend, keeping sensitive data secure within the user's infrastructure, and it interfaces smoothly with other projects such as AgentGate and FormBridge to facilitate a comprehensive lifecycle management solution for AI agents.