Show HN: Sqry – semantic code search using AST and call graphs (sqry.dev)

Sqry has introduced a new semantic code search tool that utilizes abstract syntax trees (AST) and call graphs to revolutionize the way developers interact with their codebases. Unlike traditional text-based search that may yield imprecise results, sqry processes code like a compiler, creating a detailed graph of symbols and relationships. This enables users to perform structural queries that return exact results within milliseconds, significantly enhancing the efficiency of code analysis and debugging. The integration with AI assistants such as Claude, Codex, and Gemini allows for precise querying capabilities rather than relying on probabilistic guessing. Sqry supports 35 programming languages, with full relation tracking for 28 of them. A standout feature is its ability to index the entire Linux kernel, comprising 28 million lines of code, in under two minutes, resulting in a graph with over 11 million nodes. Additionally, sqry's natural language processing capabilities allow users to query in plain English, with an impressive 99.75% intent accuracy achieved by a locally running machine learning classifier. The focus on security, with verifiable releases and comprehensive audits, further positions sqry as a trustworthy tool for developers.