# Context Engine MCP server

> Lets AI coding assistants run semantic code search, symbol graph navigation, and persistent memory queries over an indexed codebase.

- Repository: https://github.com/Context-Engine-AI/Context-Engine
- Page: https://hostedmcp.org/servers/context-engine-ai/context-engine
- Transport: —
- Stars: 401
- License: MIT

## FAQ

### What can an AI assistant do with the Context Engine MCP tools?

It can run semantic code search, explain how code works across files with citations, navigate a symbol graph, and store and recall memory across sessions.

### How do I connect my codebase to Context Engine from the terminal?

Install the MCP bridge globally with npm, then run ctxce connect with your API key and workspace path, optionally as a background daemon.

### Which tool should the assistant call first for codebase questions?

It should call repo_search first, before ls, find, Grep, or Read, and use context_answer for how/why questions spanning several files.

### Can it store knowledge across sessions?

Yes, the skills teach memory_store and memory_find for persistent context across sessions.

### Does it support searching git history?

Yes, tools like search_commits_for and change_history_for_path search git history.