# How do I stop re-explaining my project to AI agents?

Put stable instructions in CLAUDE.md or AGENTS.md, then use StremAI for learned project memory. The instruction file handles commands and conventions; StremAI stores decisions, pitfalls, handoffs, and recurring context that connected agents can recall later.

Verified with Claude Code and Claude Desktop; Cursor, Codex, Windsurf, OpenClaw, and other MCP clients are supported or in active verification.

## Use two layers

Use instruction files for rules you write by hand. Use StremAI for lessons agents discover while working.

## Make the first memory concrete

Store "our staging database resets every Sunday," start a new session, and ask what the agent knows about staging.

## When you do not need StremAI

Skip it if one tool and one repository already work without repeated context loss.

## Start

Connect your first agent: https://www.stremai.com/getting-started