SigMap builds a deterministic, auditable signature-and-evidence map of your codebase without LLM calls or embeddings. It allows AI agents, CI, and reviewers to trust and verify code references, significantly reducing token usage and improving task success rates.

sigmap: Builds a deterministic, auditable…

Features

01Deterministic, Auditable, and Verifiable Code Mapping: Generates byte-stable code maps without LLM inference, ensuring consistent and trustworthy output for AI.

02Zero Dependencies and Offline Functionality: Runs locally with no external dependencies, vector databases, or hosted services, ensuring full privacy and offline use.

03AI Answer Grounding and Fabrication Detection (`sigmap verify`): Flags fabricated files, imports, and symbols in AI answers by tracing them to real line anchors in the codebase.

04Broad Language Support: Supports 35+ programming languages, including TypeScript, Python, Go, Rust, Java, and R.

05Model-Agnostic and Agent/IDE Integrations: Works with cloud LLMs (GPT-4, Claude), local LLMs (Ollama), open-source agents (Aider), and popular IDEs (VS Code, JetBrains, Neovim).