TempoGraph is a code graph context engine that parses codebases using tree-sitter (supporting 170+ languages), builds structural dependency graphs, and exposes 24 MCP tools to give AI coding agents exactly the right files before making changes. It uses triple search (symbol, semantic, and keyword) with a single prepare_context call that adaptively selects context type within a token budget. Benchmarked at +27% F1 improvement on change-localization tasks, it is the only code context tool publishing statistically significant retrieval benchmarks.
Visit TempoGraph's official website for product details and getting started.
Comprehensive API reference and usage guides for TempoGraph.
Join discussions, ask questions, and share experiences with other TempoGraph users.