OpenReflex MCP server

OpenReflex is a memory helper for AI coding agents. It quietly records how your coding tasks go, then reminds your AI of what worked before when a similar task comes up. It is handy for developers who use Claude Code, Codex, Cursor or OpenCode and keep repeating the same fixes.

What this MCP server does

You work normally in your coding agent. Behind the scenes, OpenReflex watches the steps of each task and saves a short, private summary of what happened. Before a new task, it looks for similar past tasks and quietly adds the useful bits to your AI's context. While you work, it notices when the AI is repeating the same failing command or going in circles, and it suggests whether to continue, switch approach, or stop and ask you. After the task, it learns from the result.

Click to zoom

What you can do with it

• Reuse a fix that worked on a similar task before
• Get a warning when your AI is stuck in a loop
• Ask whether to continue, pivot or stop on a tricky task
• Preview what past experience a new task would receive
• See what has been learned about your project
• Explain why a recommendation was made
• Delete the memory of one past task

Try asking your AI

• “Fix the login redirect bug”
• “Why did you recommend pivoting just now?”
• “Show me what you remember about this project”
• “Forget the memory of the last task”

What it gives back to you

You get short, plain answers in the chat. Sometimes it is a list of similar past tasks with their lessons, sometimes a one-line suggestion like continue, pivot or stop, and sometimes an explanation of why. It can also show a timeline of the decisions made during a task.

Before you start

Who it's for

Developers and engineers who use AI coding agents daily and want their agent to remember what worked on the same project.