Supported File Types

Fresh

Code — Tree-sitter AST (Local, No LLM)

These files are processed entirely locally via tree-sitter. No network calls, no API costs.

Language	Extensions	Extracts
Python	.py	Classes, functions, imports, call graphs, docstrings
JavaScript	.js, .mjs, .cjs	Functions, classes, imports, exports
TypeScript	.ts, .tsx	Same as JS + type annotations
Go	.go	Functions, structs, interfaces, imports
Rust	.rs	Structs, enums, functions, traits, use statements
Java	.java	Classes, methods, imports
C	.c, .h	Functions, structs, #include statements
C++	.cpp, .hpp, .cc	Classes, methods, templates
Ruby	.rb	Classes, methods, requires
C#	.cs	Classes, methods, namespaces
Kotlin	.kt	Classes, functions, imports
Scala	.scala	Classes, objects, imports
PHP	.php	Classes, functions, use statements
Swift	.swift	Classes, structs, functions
Lua	.lua	Functions, requires
Zig	.zig	Functions, structs
PowerShell	.ps1	Functions, cmdlets
Elixir	.ex, .exs	Modules, functions, aliases
Objective-C	.m, .mm	Classes, methods
Julia	.jl	Functions, modules
Verilog	.v	Modules, ports
SystemVerilog	.sv	Modules, interfaces
Vue	.vue	Components, props, emits
Svelte	.svelte	Components, props
Dart	.dart	Classes, functions, imports

Documents — LLM Extraction (Sends to API)

Format	Extension	Notes
Markdown	.md	Concept and relationship extraction
Plain Text	.txt	Concept extraction
reStructuredText	.rst	Concept extraction
Word Document	.docx	Full text extraction
Excel	.xlsx	Table and data extraction
PDF	.pdf	Full text + citation extraction

API Cost Note

Document extraction sends file content to your platform's LLM API. Large PDF libraries can generate significant API costs. Use .graphifyignore to exclude docs you don't need in the graph.

Media — Vision and Transcription (Local)

Type	Extensions	Processing
Images	.png, .jpg, .jpeg, .webp, .gif	Claude vision model — extracts concepts from diagrams, screenshots, whiteboard photos
Video	.mp4, .mov, .webm	faster-whisper transcription (local) + LLM concept extraction
Audio	.mp3, .wav	faster-whisper transcription (local) + LLM concept extraction
YouTube	URL	Downloads transcript via yt-dlp + LLM concept extraction

Privacy

Video and audio transcription runs locally via faster-whisper. No audio is sent to any external server.

What Gets Excluded Automatically

Graphify excludes these by default even without a .graphifyignore:

• .git/
• node_modules/
• __pycache__/
• .venv/, venv/
• dist/, build/
• graphify-out/ (prevents graph from indexing itself)
• Binary files (executables, compiled artifacts)
• Files larger than the configured max size (default 10MB)