A step-by-step guide to building a coding assistant using Weaviate's built-in MCP server (available since v1.37.1) to enable RAG over a private codebase and documentation. The setup involves enabling the MCP server via environment variables, designing two Weaviate collections (code chunks and doc chunks) with appropriate tokenization, ingesting code using AST-aware chunking (Python's ast module or tree-sitter) and docs using heading-based chunking, then connecting Claude Code, Cursor, and VS Code to the same Weaviate endpoint. Hybrid search (BM25 + vector) is highlighted as key for code retrieval since it handles both exact identifier matching and semantic intent queries. The post also covers multi-tenancy for multi-repo organizations, RBAC auth, agent write-back for long-term memory, and embedding model choices.

10m read timeFrom weaviate.io
Post cover image
Table of contents
Why your coding assistant needs more than its training data ​Why Weaviate MCP fits this job ​Step 1: Run Weaviate with MCP enabled ​Step 2: Design the schema ​Step 3: Chunk and ingest the codebase ​Step 4: Chunk and ingest documentation ​Step 5: Connect Claude Code, Cursor, and VS Code ​Try it out ​Going further ​Summary ​Ready to start building? ​Don't want to miss another blog post?

Sort: