All activity
AgentLens is a self-hosted observability platform built specifically for AI agents. Unlike LangSmith (cloud-only) and Langfuse (LLM-focused),AgentLens understands agent topology โ tool calls, handoffs, sub-agent spawns, and decision trees.
๐ Agent Topology Graph
โช Time-Travel Replay
๐ Trace Comparison
๐ฐ Cost Tracking
๐จ Alerting
๐ก Live Streaming
๐ OTel Ingestion
One-command deploy: docker run -p 3000:3000 tranhoangtu/agentlens-observe
AgentLens-ObserveOpen-source observability for AI agents โ self-hosted
Tu Tran Hoangleft a comment
Hey PH! ๐ I built AgentLens because I kept losing hours debugging multi-agent AI systems. The existing tools either lock your data in the cloud or don't understand agent-specific concepts like tool handoffs and sub-agent spawns. AgentLens is fully self-hosted (Docker one-liner), free forever (MIT), and designed from day one for agents โ not just LLM calls. The topology graph is the feature I'm...
AgentLens-ObserveOpen-source observability for AI agents โ self-hosted
Tu Tran Hoangleft a comment
mcpman is a CLI tool that installs, manages, and health-checks Model Context Protocol (MCP) servers across all your AI coding clients โ Claude Desktop, Cursor, VS Code, and Windsurf โ from a single command. No more editing JSON configs by hand. No more guessing if a server is healthy. - mcpman install โ resolves from npm, Smithery, or GitHub - mcpman list โ shows all servers with health status...
mcpmanThe package manager for MCP servers
The package manager for MCP servers โ install, manage & monitor across Claude Desktop, Cursor, VS Code, Windsurf - tranhoangtu-it/mcpman
mcpmanThe package manager for MCP servers