Socraticode

About

Socraticode provides AI assistants with instant, automated knowledge of entire codebases at scale, offering deep semantic understanding without configuration. It's a production-ready, battle-tested MCP server designed for large repositories, featuring hybrid semantic and lexical code search, polyglot dependency graphs, and searchable infrastructure/API/database artifacts. Running locally with Docker, it ensures full privacy, automates indexing with resumable checkpoints, and keeps the index updated in real-time. It's built on Qdrant, supports various embedding providers (local Ollama, OpenAI, Google Gemini), and is measurably more efficient than grep-based exploration for AI agents.

Key Features

• Fully private and local by default, powered by Docker-managed Qdrant and Ollama.
• Polyglot code dependency graphs with static analysis and circular-dependency visualization.
• Hybrid code search (semantic + BM25, RRF-fused) for conceptual and exact queries.
• Batched, resumable, and incremental indexing with live file watching for automatic updates.
• AST-aware chunking for higher-quality results, splitting files at function/class boundaries.
• 3 GitHub stars

Use Cases

• Quickly understanding existing codebases across different programming languages and frameworks.
• Automating code intelligence for AI development workflows, reducing context and tool calls.
• Equipping AI assistants with deep semantic understanding of large and complex codebases.