Perfetto MCP Rs, developed by Tooluse Labs, is a free Model Context Protocol (MCP) server designed to empower Large Language Models (LLMs) with Perfetto trace analysis capabilities. Built with Rust for the mcp platform, this tool connects directly to Perfetto trace files and executes PerfettoSQL queries. It automatically manages its trace processing engine, making it a valuable asset for AI agents and developers. Perfetto MCP Rs provides a bridge between complex performance data and natural language queries, simplifying debugging and optimization workflows across various clients.

What does Perfetto MCP Rs connect to?

Perfetto MCP Rs serves as an intermediary, linking AI clients to Perfetto trace files such as .pftrace, .perfetto-trace, and .bin formats. It interfaces with a core dependency, trace_processor_shell, which the server automatically downloads during its initial execution. This connection enables the server to run sophisticated PerfettoSQL queries against the loaded trace data. Its primary function is to interpret and process detailed system performance traces, making the underlying data accessible for analysis by connected LLMs and agentic clients.

Core features and tools

The server exposes several specific callable tools for trace interaction and analysis. These include load_trace for importing trace files, chrome_scroll_jank_summary for specialized performance diagnostics, and execute_sql to run custom PerfettoSQL queries. For data exploration, list_stdlib_modules helps discover available standard library functions, while list_tables and list_table_structure enable clients to understand the schema of loaded traces. These tools allow LLMs to perform detailed investigations into Perfetto trace data efficiently.

Who is it for?

Perfetto MCP Rs targets a diverse audience, including Large Language Models and agentic MCP clients like Claude Code and Cursor. Android developers debugging performance issues, performance engineers analyzing complex system traces, and QA teams investigating application responsiveness problems will find it beneficial. It helps these users translate natural language questions into PerfettoSQL queries for deeper insights. This is not the right fit if you require a trace analysis tool that does not integrate with AI or agentic clients.

How to get started

No credentials are required for server access. The transport protocol used for communication between the client and server is Standard I/O (stdio), which relies on JSON-RPC 2.0 messages.

Use cases

• Loading a Perfetto trace, like "scroll_jank.pftrace", and querying in plain language for top scroll-jank causes, which activates the chrome_scroll_jank_summary tool.
• General Perfetto trace analysis by pointing an MCP client at a trace file and querying it with natural language, allowing the server to translate into PerfettoSQL.
• Executing custom PerfettoSQL queries using the execute_sql tool, potentially after discovering the trace schema with list_tables and list_table_structure.

Perfetto MCP Rs by Tooluse Labs provides a valuable bridge for AI-driven Perfetto trace analysis, simplifying complex performance debugging. Its ability to process traces and execute PerfettoSQL queries via natural language makes it highly useful for developers and performance engineers. The server integrates with various MCP clients, enhancing diagnostic workflows. Its free license and automatic dependency management offer accessibility. Try analyzing a trace file with an MCP client and asking a natural language question about performance.