Vectr MCP server

Vectr is a small helper program that runs on your own computer and gives your AI code editor a better memory of your project. Instead of re-reading the same files and hunting for the same function over and over, your AI can ask Vectr to find code by description and to recall notes it saved earlier. It is handy for anyone working in a large or unfamiliar codebase, or for work you come back to after a break.

What this MCP server does

You ask your AI something about your project, like where a certain behaviour lives. Your AI sends that request to Vectr, which is running in the background on your computer. Vectr looks through its index of your files, which it built by reading your code and understanding its structure, and finds the most relevant pieces. It sends those back to your AI, which then explains or uses them in the chat. Vectr can also save notes your AI writes down and bring them back later, even after you close and reopen your editor.

Click to zoom

What you can do with it

• Search your codebase by describing what the code does, not just by exact names
• Look up a function, class, or symbol and get the whole thing back
• Trace how functions call each other or how a route is wired up
• Save a finding as a note so your AI can recall it later
• Recall saved notes by meaning, even with different wording
• Check the status of indexing or the memory store
• Have relevant notes appear automatically in editors that support session hooks

Try asking your AI

• “Find the function that validates JWT tokens in this project”
• “Where is the user signup route defined and what does it call?”
• “Remember that the payment retry logic lives in billing/retry.py and uses exponential backoff”
• “What did we figure out earlier about the caching layer?”

What it gives back to you

You get answers in the chat, usually with the relevant code snippets or a summary of what was found. For a search, your AI shows the matching function or block. For a recall, it brings back the note you saved, in the same words. For a trace, it shows the chain of calls or routes. In editors with session hooks, some notes may appear on their own without you asking.

Before you start

Who it's for

Developers and anyone using an AI code editor on a large or unfamiliar codebase, especially work they resume across sessions.