Total Agent Memory MCP server

Total Agent Memory is a small memory program that runs on your own computer. It lets your AI coding helper remember things from earlier chats, like why you picked a certain database or how you fixed a bug. It is handy if you use tools like Claude Code, Codex, Cursor or any other app that speaks MCP.

What this MCP server does

You tell your AI something worth remembering, like a decision or a fix. The AI passes that note to this helper program. The helper writes it into a memory folder on your computer, using a small database plus a kind of search index. Later, when you ask a question, the AI asks the helper to search that folder. The helper finds the closest matches and hands them back, and the AI uses them to answer you.

Click to zoom

What you can do with it

• Save decisions, fixes, facts and errors from a session
• Recall past decisions by asking a plain question
• Keep project conventions in one place across sessions
• Track when a fact changed and retire the old value
• Read your codebase so the AI knows its structure
• Summarise a session so you can pick it up later
• Search across projects for similar past problems

Try asking your AI

• “Remember that we chose PostgreSQL for billing because of row-level security”
• “Which database did we choose for billing?”
• “What did we decide about the login flow last week?”
• “Find any past notes about the payment bug we fixed”

What it gives back to you

You get answers in the chat, drawn from your saved notes. The helper can return lists of matching memories, short summaries, or the exact note you saved earlier. If you ask it to save something, it confirms that the note was stored. You can also open a small dashboard to browse what is in memory.

Before you start

Who it's for

Developers and researchers who use AI coding helpers every day and are tired of repeating the same context.