robo-cortex MCP server

robo-cortex is a small memory keeper for coding projects. It stores lessons your AI learns, like a fix that worked or a decision you made, and ties each one to the exact version of the code it came from. It is handy for developers and anyone who uses an AI coding assistant on a project that lives in git.

What this MCP server does

You ask your AI to fix a bug or work on a feature. The AI uses this helper to check the memory store for anything related to the task. If it finds a lesson, it reads it and uses it. When the AI learns something new, it can save that as a memory too. Later, if the code that a memory points to changes, the memory gets flagged so you know to review it.

Click to zoom

What you can do with it

• Save a lesson your AI learned so it is not forgotten
• Look up past lessons and decisions for a task
• Search memories by keyword
• Attach proof like test output or a command to a memory
• Change a memory's status, for example mark it as no longer true
• See which memories are affected by recent code changes
• Read the full details of a single memory

Try asking your AI

• “Look up any lessons we have about the clipboard feature”
• “Save a memory that Safari 14 breaks CSS grid gap, use flexbox instead”
• “Which memories are affected by the changes in the last commit”
• “Mark memory 7 as superseded by memory 12 because the new pattern handles edge cases”

What it gives back to you

You get back a list of matching memories with a short statement, a score, and a flag if any need review. If you ask for one memory, you get its full details including status, evidence, and linked files. When you save or change a memory, you get a confirmation with the memory's id. Everything appears right in your chat.

Before you start

Who it's for

Developers and technical teams who use AI coding assistants on git projects and want the AI to remember what it learned.