LLMxRay: Open-Source Local Observability for LLM Traffic, Launches With One Command
GuruCsharp · reddit · 2026-10-04
- Developer releases LLMxRay, an open-source, 100% local observability and analytics engine for self-hosted/local LLM setups (Ollama, local inference servers, agentic frameworks), startable with npx llmxray — no data leaves your machine.
- Key features:
- Tool & function calling telemetry: intercept agent tool calls in real time, inspect payloads, frequencies and error rates, and analyze multi-step loops to pinpoint stalls.
- Request & latency analytics: TTFT, prefill and decode breakdowns, token usage, throughput and per-model utilization trends.
- Privacy-first with a local web dashboard at localhost:3000 for instant visual debugging.
- Repo: github.com/LogneBudo/llmxray; docs at lognebudo.github.io/llmxray.
More from coding & agent
- Blogger delegates visa applications entirely to AI agents, 3 approved so far — AlchainHust · 2026-10-04
- theo explains how he juggles 6 Claude subs and 3 Codex subs — 'sorry if it gets you banned' — 0xkarasy · 2026-10-04
- Super Mario 64 gets ported into Halo, the latest in AI coding-powered game mashups — mark_k · 2026-10-04
- Astra one-shots a zero-asset custom-engine game, developer impressed — Dimillian · 2026-10-04
- Solana AI Agent Project 'Poly' Outlines Q4 Roadmap: Own Models, Agent Toolkit, Tokenomics — DionysianAgent · 2026-10-04
- How to Build a Hiring-Worthy RAG Engineer Portfolio Project in One Month — ashishllm · 2026-10-04