Your personal memory, across sessions, agents, and devices.
Product decisions, implementation notes, and practical thinking about personal memory across agents.
Articles marked Earlier-version content are preserved for historical context and may not describe the current memU architecture or setup.
agentmemory captures coding agent sessions into a local server with no LLM calls by default. What it stores and injects, what it costs, and how it compares.
Android Studio Rabbit 2 runs Claude Agent, Codex and 40+ other ACP coding agents. How setup, billing and project context work, and what each agent remembers.
How Claude Code compaction works. When auto-compact triggers, what survives /compact, how to steer the summary, and what CLAUDE.md and memory reload after.
claude-mem records Claude Code tool calls, has a second model compress them, and loads past work at session start. Here is what it stores and costs.
Google Cloud Agent Compute Engine runs each agent session in its own microVM with no credentials inside. What persists, what it costs, and when it ships.
The main types of AI agent memory, and how Mem0, Letta, Zep, LangMem, TencentDB Agent Memory, claude-mem and memU store it, with a guide to choosing.
Claude Duet is Anthropic's proactive agent preview. How often it checks your apps, what it remembers, how to reset it, and how it compares with OpenAI dots.
Meta says Muse Wonder's Active Memory carries user context across conversations. What its docs confirm and what its MemoryBench test really measures.
DeepSeek's Super-agent prototype, dsh-superagent, is a proactive coding agent for one GitHub repo. How it works and how it keeps team context.
A Microsoft paper pairs episodic records with relational tuples and a consolidation step for long-term agent memory. How it works and what its scores mean.
Claude Code, Codex and Cursor each keep their own memory. How to share one AGENTS.md across all three, what imports copy, and where learned memory stays put.
Databricks Agent Memory, in Beta on Agent Bricks, keeps sessions and distilled memories in Lakebase. How it works, what it costs, and its limits.