profiling

6 servers · 73★ total

Runtime evidence that helps agents trace, profile, and burn down hotspots in application and native code, GPU kernels, and inference stacks.

Perfetto trace analysis via MCP — PerfettoSQL queries plus dedicated Chrome tools for scroll jank, page loads, and main-thread hotspots

CLI-first TPU/GPU profile analysis for AI agents (xprof/openxla) — roofline, HLO, Pallas kernel/LLO tools; also ships an MCP server

A live terminal dashboard for a SereneDB server: storage, memory, sessions, per-thread CPU and a perf-backed profile, cheap enough to leave open.

LLM-ready JVM profiler — turns a JFR recording into a compact, source-attributed hot-path & allocation summary that AI coding agents can act on. In-process agent, MCP server, and JMH plugin; library + CLI.

Bridge React’s Profiler API to Model Context Protocol (MCP): a local companion server + browser collector so agents answer performance questions from real render data.