Cicada
Equips AI coding assistants with deep code intelligence, enhancing contextual understanding and significantly reducing token usage.
About
CICADA is a Model Context Protocol (MCP) server that provides AI coding assistants with comprehensive code intelligence. By indexing codebases using tree-sitter AST parsing, it offers instant access to modules, functions, call sites, and PR attribution, initially for Elixir projects. This deep understanding allows AI agents to operate with significantly fewer tokens and faster response times, providing context-rich information for tasks like code search, usage tracking, and historical analysis. It also incorporates advanced AI-powered semantic search using BERT and GloVe models for superior keyword extraction and concept discovery.
Key Features
- 8 GitHub stars
- AST-aware code search with full signatures and types
- Intelligent call site tracking with alias resolution
- AI-powered keyword extraction and semantic search
- Semantic module analysis and dependency mapping
- PR attribution and historical code review context
Use Cases
- Provide AI coding assistants with precise code context for improved performance.
- Efficiently search for function definitions, call sites, and module dependencies.
- Attribute code changes to specific pull requests and access review comments.
