Build Up Claude's Memory Over Time with CLAUDE.md
Claude Code's memory lives in plain text files — your global ~/.claude/CLAUDE.md and each project's .claude/CLAUDE.md. Building these files up intentionally over time is one of the most effective ways to improve every session.
For project-specific preferences, ask Claude to write rules directly to CLAUDE.md:
Add a rule to CLAUDE.md: always use the service layer for business logic —
controllers should only handle HTTP.
Over time, your CLAUDE.md files accumulate exactly how you like to work — your test framework preferences, naming conventions, commit message format, when to ask vs when to proceed. Claude loads these at the start of every session, so you stop re-explaining the same things.
The global file is the right place for universal preferences:
## My Preferences
- I'm a senior developer — skip basic explanations
- Use conventional commits (feat:, fix:, chore:)
- Prefer composition over inheritance
- Always write tests for new functionality
Project-level CLAUDE.md is for project-specific rules that would be noise everywhere else.
For quickly dropping notes mid-session, use the # shortcut (covered in the dedicated shortcuts tip). For building memory from corrections, see the "Auto-Update CLAUDE.md Every Time You Correct Claude" tip — together these habits turn every session into an investment in every future one.
The most valuable CLAUDE.md is the one that has grown through real corrections — every mistake you fix and capture makes every future session sharper.
Log in to leave a comment.
When Claude writes error messages, button labels, validation text, or onboarding flows, it defaults to generic developer-speak. Add a "Users" section to your CLAUDE.md describing who your actual users are — their technical level, industry jargon, and what they care about — so Claude writes copy that makes sense to THEM, not to developers.
Use the --agent flag with custom markdown files in .claude/agents/ to launch purpose-built Claude sessions with restricted tools and scoped system prompts.
Every project has traps — the billing module that silently fails if you forget to queue the job, the legacy table with column names that don't match the model, the config value that must be set before tests run. Document these gotchas in your CLAUDE.md so Claude avoids the same mistakes your team spent days debugging.