Letta
Episodic memory for agents that think across sessions
AI Agents / Memory & Evals · Apache-2.0 · 25.2k
Reviewed 2026-08-19 · last push 2026-10-18 · repo checked 2026-10-21
What's Good
Letta (formerly MemGPT) pioneered the operating-system approach to agent memory. Instead of stuffing everything into context, it pages older information in and out of the working window, letting agents maintain coherent threads across sessions that would otherwise exceed context limits.
The Catch
More complex to reason about than simple vector retrieval. The paging mechanism adds latency and orchestration overhead. Documentation assumes understanding of the original MemGPT paper.
Verdict
The memory architecture for agents that need to think long-term.
Embed
Maintainer? Add the badge to your README. It is free, links to this review, and there is no paid version.
Markdown
[](https://stackgems.com/gems/letta)
HTML
<a href="https://stackgems.com/gems/letta"><img src="https://stackgems.com/badge/letta.svg" alt="Letta reviewed on Stack Gems" height="20"></a>
Related
AI Agents →- Mem0Memory infrastructure for AI agents that actually persists
- LangfuseOpen-source LLM observability you can self-host
- PromptfooMIT LLM evals & red-teaming CLI; Community free, 10k probes/mo
- DeepEvalApache-2.0 pytest-style LLM/agent evals; Confident AI is the paid cloud
- GraphitiApache-2.0 temporal knowledge graphs for agent memory; Zep's OSS engine