Letta Code
A stateful agent harness whose context, memory, and skills can evolve across sessions. It supports local CLI interaction and optional remote or messaging access.
Utility: Knowledge and memory; kind: agent.
Stars: 3k+; Starred in September: 319.
Metrics updated: 31 Oct 2026. Profile reviewed: 31 Oct 2026.
Utilities
- Maintain an agent's working memory across longer-running interactions.
- Inspect and revise learned context or skills through the documented interface.
Who it is for
- Technical users experimenting with persistent rather than disposable agent sessions.
Requirements
- Install the npm package or supported Python wrapper and configure a model.
- Remote-computer and secrets features require the documented Letta sign-in.
Limitations
- Automatically rewritten memory or skills can drift and need inspection.
- Platform support differs by installation route; check the current wheel and operating-system requirements.
Cost and privacy notes
- Model-provider usage and hosted services may have separate costs.
- Persistent identity and remote-computer access increase the need to review stored context and permissions.
Illustrative example (untested)
Illustrative, not tested: ask an agent to retain an approved project convention, then inspect its memory before the next session.
License
Apache-2.0; reviewed the current letta-code repository LICENSE.