Graphify: Giving Your AI Coding Assistant a Persistent Map of Your Codebase

Grep reads your codebase. Graphify understands it. I got tired of watching my AI coding assistant re-read the same 40 files every single session just to "remember" how the codebase fits together. So I tried Graphify - an open-source CLI that turns any repo into an actual knowledge graph your assistant can traverse, instead of grepping or fuzzy-matching chunks from scratch each time. One command: /graphify . It parses your code locally with 36 tree-sitter grammars (no LLM, nothing leaves your machine), then builds a graph of functions, classes, and files linked by real relationships - calls, imports, inherits. The graph persists, so the assistant opens already knowing the architecture. What stood out: • Every edge is tagged EXTRACTED (found in code) or INFERRED (reasoned) - no black-box similarity scores • Auto-detects "god nodes" and module communities via Leiden clustering • Ships as a skill for Claude Code, Cursor, Copilot, Codex, Gemini CLI, and 15+ other assistants If your assistant is starting from zero every session too, worth a look. 🔗 graphify.com | https://lnkd.in/gpNPp2fv #AI #DeveloperTools #OpenSource #KnowledgeGraph #AICoding #SoftwareEngineering #DevTools #RAG #CodeIntelligence #BuildInPublic

Confirmed accurate (verified against Graphify's own docs earlier in this conversation):

• Graphify is real: an open-source /graphify skill for Claude Code, Cursor, Codex, and Gemini CLI that does local deterministic AST parsing with no vector store — matches "open-source CLI" here. GitHub
• Code is parsed with tree-sitter AST — deterministic, no LLM, nothing leaves your machine — matches "parses your code locally... no LLM, nothing leaves your machine." GitHub
• Graphify ships grammars for thirty-six programming languages via Tree-sitter, operating entirely on the local machine with zero API calls — matches "36 tree-sitter grammars" exactly. Learn AI
• Relationships are extracted as calls, imports, and inheritance edges, and every relationship is tagged EXTRACTED (found directly in source, confidence 1.0) or INFERRED (a reasonable inference with a confidence score) — with AMBIGUOUS as a third tag for uncertain cases the model flags for review. The post's "EXTRACTED or INFERRED" framing is accurate but incomplete — it omits AMBIGUOUS, the third tag Graphify actually uses for low-confidence relationships. Minor simplification, not a misstatement. GitHubGitHub
• Communities are found using the Leiden algorithm, a graph-clustering method that groups nodes by edge density, and other write-ups describe the tool surfacing "god nodes" — matches the "god nodes and module communities via Leiden clustering" claim. Emelia
• Graphify installs as a skill in 17 assistants, including Claude Code, Cursor, GitHub Copilot, Codex, Gemini CLI, and Aider — supports "Claude Code, Cursor, Copilot, Codex, Gemini CLI, and 15+ other assistants" (17 total is consistent with "5 named + 15+ more" as a reasonable rounding). Graphify
• The persistence claim is accurate: the graph is written to disk and can be committed to Git and refreshed incrementally, so the assistant doesn't rebuild its understanding from scratch each session.

No factual errors found. This is a shorter, slightly more conservative version of the earlier Graphify post — it drops the star-count claim entirely, which sidesteps the one figure that showed inconsistency across trackers in my earlier check.