Sqry by Verivus Oss is a Model Context Protocol (MCP) server for AI assistants and a Language Server Protocol (LSP) server for code editors. Built primarily with Rust 1.94+ and Edition 2024, it parses source code into Abstract Syntax Tree (AST)-backed symbol and relationship graphs. This server tool enhances development workflows by providing semantic code search, repository orientation, and code quality analysis. Sqry runs fully locally with no telemetry, supporting a range of AI clients and code editors.

What does Sqry connect to?

Sqry primarily supports the stdio transport protocol for local MCP and LSP operations. It integrates with compatible AI clients including Claude, Codex, Gemini, Cursor, Windsurf, Claude Code, and ChatGPT. For code editors, Sqry offers LSP support for VS Code, Neovim, and Helix. Future plans include support for remote MCP via HTTPS Streaming or SSE, extending its reach beyond local environments for more complex setups.

Core features and tools

Sqry functions as a semantic code search engine, parsing source code into AST-backed symbol and relationship graphs. It exposes a varying number of callable tools, depending on the mode. The standalone sqry-mcp offers 39 tools, while the daemon-hosted MCP provides a 17-tool subset. Other sources indicate 38 tools for the MCP server and 37 for the "--no-daemon" mode, with 16 tools for daemon mode. The authoritative catalog is available via "tools/list" or "sqry-mcp --list-tools".

Who is it for?

Sqry is designed for developers and AI agents/assistants looking to deepen their understanding of codebases and automate code analysis tasks. It aids in navigating complex projects, identifying structural relationships, and ensuring code quality. Developers can integrate it into their daily workflows for powerful semantic queries and repository insights. This is not the right fit if you require remote Model Context Protocol capabilities today.

How to get started

No credentials are required. It primarily uses the stdio transport protocol for local operations.

Use cases

• Semantic code search: Answering structural queries like "who calls this function" or "every implementation of this trait". For example, "$ sqry query callers:authenticate".
• Repository orientation: Providing a "one-shot repository orientation" report that identifies load-bearing hubs, package subsystems, complexity hotspots, and potential issues.
• Code quality analysis: Detecting unused symbols or dead code, for instance, "$ sqry unused --lang rust". It can also be integrated into CI for identifying dead code, cycles, and duplicates on every pull request.

Sqry is a powerful, locally-run tool that significantly enhances code understanding and analysis for developers and AI assistants. Its ability to parse code into AST-backed graphs enables sophisticated semantic queries and quality checks. The varied distribution methods make it accessible for different development environments. Its local operation and lack of telemetry ensure privacy.