Is a manual Perfetto installation required to use this tool?

No, Perfetto automatically downloads and manages the `trace_processor_shell` backend on first run, eliminating the need for any manual Perfetto installation.

Can I use Perfetto for custom queries on non-Chrome traces?

Yes, you can use PerfettoSQL to discover schema with `list_tables` and `list_table_structure`, then execute custom queries to extract any data from your Perfetto trace files.

Which LLMs and clients does Perfetto integrate with?

It offers seamless integration with LLM-powered Model Context Protocol (MCP) clients such as Claude Code, Codex, Claude Desktop, and Cursor.

What types of performance issues can Perfetto help analyze?

Perfetto provides specialized tools for deep analysis of Chrome performance traces, including scroll jank, page loads, main thread hotspots, and startup summaries.

Perfetto

Name: Perfetto
Author: 0xZOne

by0xZOne

•

开发者工具

分析与监控

数据科学与机器学习

Enables Large Language Models to analyze Perfetto trace files, allowing queries in PerfettoSQL through any Model Context Protocol client.

This tool provides a Model Context Protocol (MCP) server designed to enable Large Language Models (LLMs) to interactively analyze Perfetto trace files. It allows users to point an MCP client, such as Claude Code or Codex, at `.pftrace` files and query them using PerfettoSQL. The server seamlessly manages the underlying `trace_processor_shell` dependency, downloading it automatically on first use. It is optimized for agentic MCP clients capable of multi-turn tool calls, guiding the LLM through a structured analysis workflow, from loading traces and discovering schemas to executing complex SQL queries, and offering specialized tools for Chrome traces.

主要功能

01Seamless integration with LLM-powered MCP clients like Claude Code and Codex

02Cross-platform compatibility with easy installation and upgrade mechanisms

03Specialized tools for deep analysis of Chrome performance traces (e.g., scroll jank, page load)

04Direct execution of PerfettoSQL queries for schema discovery and custom data extraction

05Automatic download and management of the `trace_processor_shell` backend

064 GitHub stars

使用案例

01Automated identification and analysis of performance bottlenecks in Chrome browser traces using LLMs

02Interactive schema discovery and querying of any Perfetto trace file via natural language

03Integrating trace analysis capabilities into LLM-driven development workflows and agents