Optimizes AI token usage by providing surgical access to codebases through intelligent indexing and retrieval.
Sponsored
TokenNuke is an intelligent code indexing server designed to dramatically reduce token consumption for AI coding agents and improve developer workflow. By leveraging tree-sitter for AST extraction and O(1) byte-offset retrieval, it allows for fetching exact function source code, saving up to 99% of tokens compared to reading entire files. It offers a comprehensive suite of 15 MCP tools, supports 10 programming languages, features incremental indexing, hybrid search (FTS5 + vector), and generates call graphs, making code interaction highly efficient and precise.
主要功能
01Incremental Indexing for fast updates
020 GitHub stars
03O(1) Symbol Retrieval with byte-offset seeking
04Remote Repository Indexing without Git
05Call Graph generation for function dependencies
06Hybrid Search combining FTS5 and vector similarity
使用案例
01Equipping AI Coding Agents with surgical code access
02Streamlining code reviews by mapping function dependencies
03Facilitating developer onboarding with semantic code search
