CodeBendKit/codeseek

Rust-powered code intelligence CLI for AI coding agents. Builds call graphs and hybrid semantic search indexes (Dense + Sparse + RRF + Reranker) across 7 languages. Ships as native MCP tools for Claude Code and Codex CLI.

What it solves

CodeSeek is a CLI tool designed to help developers and AI agents (like Claude Code) navigate and understand complex codebases. It solves the problem of finding specific logic or tracing function calls across a project without relying on simple text search, by combining semantic search with structural code analysis.

How it works

The tool uses a hybrid search pipeline to locate code symbols. It first parses source files using Tree-sitter to extract functions and classes, then creates a call graph of the project. For searching, it employs three parallel paths: dense vector search (via LanceDB), sparse keyword search (via Tantivy BM25), and graph-based name search. These results are merged using Reciprocal Rank Fusion (RRF) and then refined by a cross-encoder reranker (Qwen3-Reranker) for high precision.

Who it’s for

It is primarily for developers who want a powerful terminal-based code intelligence tool and users of AI coding agents like Claude Code or Codex, who can integrate CodeSeek as a Model Context Protocol (MCP) tool to give the agent better codebase awareness.

Highlights

  • Hybrid Search: Combines semantic embeddings, BM25 keyword matching, and call graph analysis.
  • MCP Integration: Can be registered as a tool for Claude Code and Codex to enable agentic code exploration.
  • Call Graph Analysis: Ability to trace upstream callers and downstream callees for any given symbol.
  • Incremental Indexing: Uses MD5 hashes to only re-process changed files during updates.
  • Multi-language Support: Supports Rust, Python, JavaScript, TypeScript, Go, C/C++, and Java.

Related

  • Project
    colbymchenry/codegraph

    A semantic code intelligence platform that builds a local knowledge graph of a codebase to provide AI agents with surgical context, reducing token costs and tool calls.

  • Project
    Context-Engine-AI/Context-Engine

    A semantic code search and symbol intelligence system that provides AI coding assistants with deep codebase context, memory, and navigation tools via MCP.

  • Project
    defendend/Claude-ast-index-search

    A structural, AST-aware code navigation CLI that builds a local SQLite index to help humans and AI agents navigate large repositories by exact structure rather than text matches.

  • Project
    aovestdipaperino/tokensave

    tokensave is a Rust‑based local service that builds a semantic knowledge graph of a codebase. AI coding agents query this graph via MCP tools to get symbols, relationships and snippets in one call, dramatically cutting token usage and eliminating costly file‑scanning operations. It supports 50+ languages, 12+ agent integrations, optional per‑branch indexing, crash‑resilient sync, and fully offline operation.

  • Project
    justrach/codedb

    A high-performance code intelligence server for AI agents that provides structural indexing and fast symbol search to help agents understand large codebases.