Contextual

Contextual is a local-first MCP server that gives AI coding agents persistent memory through semantic search, dependency graphs and git history.

What you can do

Provide persistent memory for AI coding agents across sessions, backed by local storage rather than a hosted service. Support semantic search over the project's code and context. Expose dependency graphs so the assistant can reason about how code is connected. Surface git history as part of the remembered context. The manifest lists no individual tools, so exact operations are not documented here.

When to use it

Worth adding when you want a coding assistant to retain project context between sessions and to look up code by meaning, dependencies or past commits instead of re-reading files each time. Best suited to local development work where keeping data on your machine matters.

Requirements

Runs as a local stdio process installed from the PyPI package contextual-engine, so a Python environment is needed. Optionally set the CONTEXTUAL_LOG_LEVEL environment variable to control logging verbosity; it defaults to INFO. No authentication is declared.