Letta Agent | Personalized Agent That Remembers and Learns

Download Log in drag_handle close dark_mode light_mode

Letta Agent<br>Build digital people whose memory, identity, and capabilities improve with experience.<br>Download for macOS Install the CLI

PERSISTENT AGENTS }}}<br>Custom agents that live forever<br>Build your own deeply personalized agents, each with their own unique identity and expertise. Supports custom skills, harness modifications, and tools.

CONTINUAL LEARNING }}}<br>Best-in-class memory & learning<br>Agents self-improve through sleep-time compute, learning without needing model updates. All context (skills, prompts, etc.) is git-tracked through MemFS.

CHANNELS }}}<br>Works where you work<br>Communicate with agents through Slack, Discord, Telegram, or custom channels. Collaboratively teach and direct agents where your team already works.

SCHEDULED TASKS }}}<br>Proactive and understands time<br>Ask agents to work proactively, monitor results, or run tasks on a schedule. Create schedules yourself or let agents set their own, and review past runs anytime.

RUN ON ANY DEVICE }}}<br>Run on remote devices<br>Connect your own device, sandbox, or VM as a remote environment. Agents can work across multiple devices simultaneously.

MULTI-MODEL MEMORY }}}<br>Bring your own model and subscription plans<br>Works with OpenAI Codex, Z.AI, and other subscription plans, or with your own API keys. Agents transfer across models without memory loss.

AI ResearchBuilding machines that learn

Letta is an AI lab building machines that learn: persistent agents with the ability to continuously learn and adapt from their own experience. Born from MemGPT at UC Berkeley's Sky Computing Lab, and backed by leaders in AI research and infrastructure.<br>Recent work<br>Continual Learning in Token Space } Sleep-time Compute: Scaling Inference-Time Compute for LLM Agents }<br>Explore our research

Research blog }<br>Context Constitution<br>Today we're releasing the Context Constitution: a set of principles governing how AI agents manage context to learn from experience.

Introducing Context Repositories: Git-based Memory for Coding Agents<br>We're introducing Context Repositories, a rebuild of how agent memory works, using programmatic context management and git-based versioning.

Continual Learning in Token Space<br>At Letta, we believe that learning in token space is the key to building AI agents that truly improve over time.

Skill Learning: Bringing Continual Learning to CLI Agents<br>Today we're releasing Skill Learning, a way to dynamically learn skills through experience.

Context-Bench: Benchmarking LLMs on Agentic Context Engineering<br>Context-Bench evaluates how well language models chain file operations, trace relationships, and manage long-horizon information retrieval.

Sleep-time Compute: AI that Dreams<br>We introduce sleep-time compute, a technique that allows agents to reason about context during idle time rather than at inference.

Compatible with your API keys and coding plans<br>Connect your own API keys and coding plans to Letta Agent for free.

Life-long agents that learn, accessible from anywhere<br>Install the desktop app}

Use in the terminal}

Build with the SDK}

Letta Agent<br>Create your own agent that remembers and learns. Try for free with your own API keys or existing coding plans.

Download for MacOS Install the CLI

Get started with Letta Code<br>close

Letta Code is a memory-first agent that can take actions on your local computer. Install in your terminal (requires Node.js 18+):<br>npm i -g @letta-ai/letta-code content_copy<br>Once installed, start Letta Code with the command below:<br>letta content_copy

Download the Letta Agent app<br>close<br>macOS (Apple Silicon) macOS (Intel) Windows Linux

Cookie Preferences<br>close<br>We use limited browser signals to understand interest in our products and improve how we share them. You can change this setting anytime.<br>Strictly necessary Needed for basic site functionality and always enabled. Marketing measurement Allows us to understand visits to this product page and related outreach performance. Done<br>Light

Dark