claude-mem, mem0, Letta, Zep, Munin: which agent memory tool fits what
Most of these tools solve a different problem. Only two compete head-on, and picking the wrong one wastes a weekend.
Read →
Blog
Notes on agent memory.
Tool comparisons, workflow posts, and the occasional opinion piece on running coding agents across Claude Code, Codex, and Cursor.
Why Claude Code forgets what you told it yesterday
Claude Code session memory is by design. The fix is an external memory layer you run yourself, not a feature request to Anthropic.
Read →CLAUDE.md, Cursor rules, mem0: three ways to give an agent a memory
Two of these are prompt prefixes. One is actual memory. Treating the first two as the third is how you end up with a 2000-line CLAUDE.md.
Read →Local-first memory for coding agents in 2026
Three of the five serious tools send your context to a cloud. The two that stay on disk do not overlap, so a solo dev can run both.
Read →From corrections to rules: teaching coding agents without more prompts
Correcting your agent the same way three times is a workflow failure, not a model bug. Promoting the correction to a rule takes two minutes.
Read →