AgentLens-Observe
Open-source observability for AI agents — self-hosted
AgentLens-Observe – Self-hosted observability platform for AI agent workflows
Summary: AgentLens-Observe is a self-hosted observability tool designed for AI agents, providing detailed insights into agent topology including tool calls, handoffs, and sub-agent spawns. It offers features like topology graphs, time-travel replay, trace comparison, cost tracking, alerting, live streaming, and OpenTelemetry ingestion.
What it does
It visualizes AI agent workflows through topology graphs and enables debugging via time-travel replay and trace comparison. The platform supports live streaming, cost tracking, alerting, and integrates OpenTelemetry data.
Who it's for
Developers and teams working with multi-agent AI systems needing detailed, self-hosted observability beyond LLM calls.
Why it matters
It solves the challenge of debugging complex AI agents by mapping their interactions and preserving data privacy through self-hosting.