Built-in memory
Best for concise user preferences and recurring workflow hints.
Comparison
remem vs built-in Claude Code and Codex memory.
Built-in memory is useful for concise preferences and stable project guidance. remem is for engineering memory that must be searchable, auditable, project-scoped, and recoverable.
Preference: "use pnpm"
Built-in memory works well
CLAUDE.md works well
Searchable bug rationale
remem is stronger
Use both for different memory shapes
Less re-explaining, better audit trail
Use cases
Choose the right memory surface.
Keep required team guidance in checked-in docs. Use built-in memory for stable preferences. Use remem when the evidence trail matters.
CLAUDE.md / AGENTS.md
Best for mandatory repository instructions that every agent must read.
remem search
Best for decisions, bug-fix rationale, timelines, and project history.
remem why
Best when you need to inspect why a memory was injected.
What remem adds
Search, audit, project scope, and APIs.
remem complements built-in memory by keeping a structured local memory store that agents and tools can query.
Searchremem search
Find historical decisions and root causes by query.
Auditremem why
Inspect why memory was visible or injected.
LocalSQLCipher
Keep project memory in a local SQLite/SQLCipher store.
APIMCP + REST
Expose memory to coding agents and local tooling.
Install
Add remem where built-in memory is too small.
Start with the host you use most, then add the other host later with `remem install --target all`.
brew install majiayu000/tap/remem
remem install --target codex
remem doctor
remem search "root cause"