Comments URL: https://news.ycombinator.com/item?id=47565649

Points: 2

# Comments: 0

I built AgentLens because debugging multi-agent systems is painful. LangSmith is cloud-only and paid. Langfuse tracks LLM calls but doesn't understand agent topology — tool calls, handoffs, decision trees.

AgentLens is a self-hosted observability platform built specifically for AI agents:

- *Topology graph* — see your agent's tool calls, LLM calls, and sub-agent spawns as an interactive DAG - *Time-travel replay* — step through an agent run frame-by-frame with a scrubber timeline - *Trace comparison* — side-by-side diff of two runs with color-coded span matching - *Cost tracking* — 27 models priced (GPT-4.1, Claude 4, Gemini 2.0, etc.) - *Live streaming* — watch spans appear in real-time via SSE - *Alerting* — anomaly detection for cost spikes, error rates, latency - *OTel ingestion* — accepts OTLP HTTP JSON, so any OTel-instrumented app works

Works with LangChain, CrewAI, AutoGen, LlamaIndex, and Google ADK.

Tech: React 19 + FastAPI + SQLite/PostgreSQL. MIT licensed. 231 tests, 100% coverage.

    docker run -p 3000:3000 tranhoangtu/agentlens-observe:0.6.0
    pip install agentlens-observe

GitHub: https://github.com/tranhoangtu-it/agentlens-observe Docs: https://agentlens-observe.pages.dev

I'd love feedback on the trace visualization approach and what features matter most for your agent debugging workflow.

Comments URL: https://news.ycombinator.com/item?id=47205382

Points: 2

# Comments: 2

  I built mcpman — a CLI tool that manages Model Context Protocol (MCP) servers across all your AI coding clients.

  The problem: If you use Claude Desktop, Cursor, VS Code, or Windsurf, you've probably edited JSON config files by hand to add MCP servers. Each client stores configs differently, there's no version pinning, and no way to know if a server
  is actually healthy.

  What mcpman does:

  - mcpman install  — resolves from npm/Smithery/GitHub, auto-configures for your installed clients
  - mcpman list — shows all servers with health status
  - mcpman doctor — runtime checks, env var validation, process spawn, MCP handshake
  - mcpman remove — clean uninstall from all clients
  - mcpman init — lockfile for reproducible team setups

  Why not Smithery CLI or mcpm.sh? Existing tools focus on registry/config. Nobody owns the operational layer — health monitoring, version pinning, cross-client sync. That's the gap mcpman fills.

  Tech: TypeScript, citty, @clack/prompts, zero daemon. ~45KB bundled. MIT licensed.

  npx mcpman install @modelcontextprotocol/server-filesystem
  npx mcpman doctor

  Feedback welcome — especially on MCP server management pain points you've hit.

Comments URL: https://news.ycombinator.com/item?id=47182793

Points: 2

# Comments: 1