I built Arbor because standard RAG-based AI tools are "architecturally blind." They see code as text snippets, but they don't understand the graph of your logic.
Arbor uses Rust and Tree-sitter to index your AST locally and maps it into a 3D Logic Forest. It’s an MCP server, so you can plug it into Claude Desktop and watch the camera "fly" to the code the AI is talking about.
Key features: 100% Local-First: Your code never leaves your machine. Sub-100ms Sync: Real-time indexing even on large repos. Cinematic Visualizer: Built in Flutter with custom GLSL shaders for a high-fidelity experience.
I'd love to hear your thoughts on using graph-based context vs. traditional vector embeddings for AI coding assistants!
holg•4w ago
OK i tried it, and in order to do so i did the needed changes for the visualizer (macOS) PR is already out.
So to me it seems the main benefit would be for interconnected projects.
anandb71•1mo ago
Arbor uses Rust and Tree-sitter to index your AST locally and maps it into a 3D Logic Forest. It’s an MCP server, so you can plug it into Claude Desktop and watch the camera "fly" to the code the AI is talking about.
Key features: 100% Local-First: Your code never leaves your machine. Sub-100ms Sync: Real-time indexing even on large repos. Cinematic Visualizer: Built in Flutter with custom GLSL shaders for a high-fidelity experience.
I'd love to hear your thoughts on using graph-based context vs. traditional vector embeddings for AI coding assistants!
holg•4w ago