gitnexus-cli

To index a local git repository for AI codebase analysis, run `npx gitnexus analyze` from the repository root to build a searchable knowledge graph of source files and relationships. This requires Node.js and npx to execute the CLI.

A repository knowledge graph maps source files and symbol relationships, enabling developers and AI assistants to reason over project architecture instead of raw files. It powers semantic search and provides context for AI copilots.

Yes, you can generate documentation and wiki files from source code by invoking the wiki generation feature via an LLM. This requires optional LLM API credentials and produces context files with optional embeddings for improved search relevance.

To check and clean up a stale codebase index, use the GitNexus CLI to perform status checks and remove corrupted or outdated indexes. The index state is written to a `.gitnexus directory` for periodic maintenance and reindexing.

No, you do not need LLM API credentials to build a codebase index. LLM credentials are only optional for generating repository wiki documentation and context files with embeddings, while standard indexing requires just Node.js and npx.

You should reindex a codebase after a large refactor to regenerate context files for AI copilots. This refreshes the knowledge graph to surface up-to-date architecture and symbol information for semantic search.