AL Perf Analyzer
A CLI, web app, MCP server, and library for analyzing Business Central .alcpuprofile files, detecting performance anti-patterns and hotspots with optional AI-powered explanations
Overview
Analyzes Business Central .alcpuprofile files to find hotspots, detect anti-patterns like CalcFields-in-loops, unfiltered FindSet calls, and N+1 query patterns, and compare profiles for regressions. Works as a CLI tool, a web app for uploading profiles, an MCP server for AI-assisted analysis, or a library you can import directly.
Two analysis levels: profile-only analysis detects call tree patterns without source code, while source-correlated analysis adds syntactic nesting validation, SetLoadFields detection, and unfiltered query detection. Optional AI explanations via the Anthropic API translate findings into plain-English summaries.
Data handling: profiles are transmitted over HTTPS, analyzed in memory, and immediately discarded. Nothing is stored server-side.