Skill Detail

Build stateful agents with long-term memory using Letta

Use Letta to create agents whose identity, tools, memory blocks, and conversation state persist across sessions for long-running assistant or application workflows.

Developer ToolsMulti-Framework
Developer Tools Multi-Framework Security Reviewed
⭐ 22.7k GitHub stars
INSTALL WITH ANY AGENT
npx skills add agentskillexchange/skills --skill build-stateful-agents-with-long-term-memory-using-letta Copy
Works best when you want a reusable capability, not another fragile one-off prompt.
At a glance
Tools required
Node.js 18+ for Letta Code or API credentials/SDKs for application integration; selected model provider
Install & setup
For local terminal workflows, install Letta Code with `npm install -g @letta-ai/letta-code` and run `letta`; for application workflows, install the TypeScript or Python Letta client and create agents with memory blocks and tools.
Author
Letta AI
Publisher
Open Source
Last updated
May 12, 2026
Quick brief

Letta provides infrastructure for stateful agents with persistent memory, tools, identities, and resumable conversation state. Use this skill when an operator or developer needs an agent to remember durable facts, resume work across sessions, or expose a memory-backed agent through a local terminal workflow or application API.

How it works

What this skill actually does

The workflow is to choose Letta Code for local terminal agent work or the Letta API/SDKs for application integration, define memory blocks and tools, create the agent, and send messages while Letta maintains state over time. The scope boundary is long-term state and memory management for agents; it is not a generic chat app, SDK reference, or model-provider wrapper.

Inputs and prerequisites: Node.js 18+ for Letta Code or API credentials/SDKs for application integration; selected model provider.

Setup notes: For local terminal workflows, install Letta Code with `npm install -g @letta-ai/letta-code` and run `letta`; for application workflows, install the TypeScript or Python Letta client and create agents with memory blocks and tools.

Source and verification boundary: use https://docs.letta.com/ as the canonical reference before running the workflow; keep commands, API calls, CLI usage, and generated outputs reviewable against that upstream source.

Framework fit: publish this as a Multi-Framework workflow only when the operator can invoke the documented toolchain directly, rather than treating the upstream project as a generic product listing.