MCPSERV.CLUB
benmyles

Glyph

MCP Server

Fast, declarative symbol extraction for AI coding agents

Stale(55)
9stars
1views
Updated Sep 18, 2025

About

Glyph is an MCP server that parses codebases with Tree‑sitter queries to produce clean, language‑agnostic symbol outlines. It supplies LLM agents with concise context across multiple files, improving code comprehension without exceeding token limits.

Capabilities

Resources
Access data sources
Tools
Execute functions
Prompts
Pre-built templates
Sampling
AI model interactions

Overview

Glyph is a Model Context Protocol (MCP) server that turns your codebase into a lightweight, structured map of symbols. By leveraging Tree‑sitter’s declarative query language, it parses source files across multiple languages and produces concise outlines of functions, classes, interfaces, constants, and more. The result is a clean context payload that keeps token usage low while giving large language models (LLMs) the precise structural information they need to reason about code, generate accurate suggestions, or refactor safely.

The server solves a common pain point for AI‑assisted development: the difficulty of feeding an LLM enough architectural insight without exhausting its context window. Traditional approaches either hand‑craft AST traversals for each language or rely on heavy, monolithic parsers that are slow and difficult to maintain. Glyph’s query‑based extraction is both fast and extensible—adding support for a new language requires only a few dozen lines of Tree‑sitter query patterns, and the same extraction engine can be reused across projects.

Key capabilities include:

  • Declarative, language‑agnostic extraction – Tree‑sitter queries replace complex code analysis logic, ensuring accuracy and maintainability.
  • Glob‑based file discovery – Developers point Glyph at familiar path patterns, and it recursively finds all matching files.
  • Configurable detail levels – Users can choose between a high‑level overview (e.g., module names) or full signatures with visibility modifiers, tailoring the output to the needs of a particular LLM prompt.
  • Unified multi‑file view – Glyph stitches together symbol information from thousands of files into a single, coherent outline that respects project boundaries.
  • MCP‑native integration – The server speaks the MCP protocol out of the box, allowing seamless attachment to Claude Code, Cursor, or any other AI assistant that supports MCP.

In practice, Glyph is invaluable for tasks such as code completion, documentation generation, or automated refactoring. An AI assistant can request the symbol map for a repository, quickly grasp the call graph, and produce context‑aware suggestions without needing to parse the entire codebase itself. Developers can also run Glyph as a standalone CLI tool for quick inspections or CI checks, but its true strength lies in the tight coupling with MCP‑enabled workflows where the server acts as a lightweight, on‑demand knowledge source.