I built CodeGraphy v2.1 - a Local Context Intelligence Platform & MCP Server with an interactive architecture graph that cuts LLM token usage by 85%

A developer released CodeGraphy v2.1, a local-first context intelligence platform and MCP server that parses a codebase's AST and imports into an interactive dependency graph. The tool ships a built-in STDIO MCP server with 11 tools for Claude Desktop and Cursor, and uses a hybrid context mode that keeps the target file in full while compressing neighboring modules to AST interface skeletons, which the developer says cuts LLM token usage by roughly 85%. It also adds blast-radius impact analysis, a Git diff overlay, and cycle detection via Tarjan's algorithm.

I built it because every time a codebase grows past 20–30 files, AI assistants either burn through token limits in 3 prompts or suffer from "lost in the middle" hallucinations. What CodeGraphy v2.1 does: - ⚡ 1-Click Project Analysis & Living Graph: Point it at a local folder — it parses AST & imports, groups modules cleanly, and builds a Pure Graphite interactive dependency graph. - 🔌 Built-in MCP Server STDIO, 11 Tools : One-click config for Claude Desktop & Cursor get architecture summary , build smart context , analyze change impact , get architecture diff , search symbols , run impacted tests . - 🧬 Smart Hybrid Context ~85% Token Budget : Keeps your target file in Full Mode while compressing connected neighbor modules down to AST interface skeletons, with provenance tracking and automatic secret scrubbing. - 💥 Blast Radius & Git Architecture Diff Overlay: Visually see 1-hop and 2-hop impact zones before you refactor, overlay uncommitted Git diffs directly on the graph, and detect cyclic dependencies via Tarjan's $O V+E $ algorithm. It's 100% local-first. I'd your honest feedback, bug reports, or a ⭐ star on GitHub if you find it useful click https://github.com/Vindettobkb/CodeGraphy-Desktop