Hicortex MCP server

Hicortex is a shared memory for AI agents. It remembers the decisions, fixes and context from your work sessions, so your AI stops asking you to explain the same things again. It is handy if you use coding assistants like Claude Code and want them to remember your project across days and machines.

What this MCP server does

You ask your AI something, like what you decided about a file last week. The AI uses this helper to search the Hicortex memory store. Hicortex looks through saved sessions, notes and learnings, and finds the parts that match. The AI then answers you using that remembered context, and can save new learnings back for next time. Each night Hicortex also reads the day's sessions, cleans up stale facts and merges duplicates, so the memory stays useful.

Click to zoom

What you can do with it

• Search your past sessions and notes for a topic
• Pull up the most recent memories from your work
• Save a new learning or note for later
• See a graph of how your memories are connected
• Update or delete a memory that is wrong or outdated
• Get a list of general lessons learned from your work
• Inject relevant memory into your coding agent's prompts automatically

Try asking your AI

• “What did we decide about the login flow last week?”
• “Save a learning: our staging server uses port 8080, not 3000.”
• “Show me the most recent memories from this project.”
• “What lessons have we learned about database migrations?”

What it gives back to you

You get answers in the chat, written in normal sentences, based on what Hicortex remembers. It can also show lists of memories, recent items, or lessons learned. When you save something, it confirms the memory was stored. There is also a dashboard and a graph view on your own machine if you want to browse the memory yourself.

Before you start

Who it's for

People who use coding assistants like Claude Code every day and are tired of re-explaining their project, especially those working across several machines or agents.