How can I deploy and interact with Grigori?

Grigori supports various deployment modes including Docker, an HTTP server, a Blazor-based interactive dashboard, and a native AOT CLI, offering flexibility for local use or remote operations.

What technology powers Grigori's code search?

Grigori leverages all-MiniLM-L6-v2 embeddings for accurate semantic understanding and the HNSW algorithm for fast, efficient vector search, providing a significant speedup over traditional search methods.

How does Grigori enhance AI assistant capabilities?

By integrating via Model Context Protocol (MCP), Grigori enables AI assistants like Claude to perform semantic searches across your codebase, allowing them to find relevant code based on meaning, not just keywords.

Grigori

Name: Grigori
Author: andres-m-rodriguez

•

Enhances AI assistant capabilities by providing deep contextual awareness of software projects through semantic code search and persistent memory.

Grigori serves as an MCP server designed to provide AI assistants with profound contextual awareness of software projects, particularly within .NET environments. It achieves this by indexing codebases once and then facilitating semantic code search using advanced embeddings and fast vector search algorithms. Seamlessly integrating with AI tools via the Model Context Protocol (MCP), Grigori supports features like language-aware code chunking and offers an interactive Blazor dashboard for managing your code index, all while striving to evolve into a comprehensive codebase intelligence system.

主な機能

01Semantic Search using all-MiniLM-L6-v2 embeddings

02Fast Vector Search with HNSW algorithm

03Model Context Protocol (MCP) Integration

04Language-Aware Code Chunking (C# support)

05Interactive Blazor Dashboard for management

062 GitHub stars

ユースケース

01Empowering AI assistants with deep codebase context for development tasks

02Performing semantic code searches by meaning across indexed projects

03Managing and indexing .NET codebases for AI-driven development workflows