← All companies

Agents / Company profile

Letta

What Letta does

Letta is an AI agent platform and research-to-software company focused on building “stateful” agents—agents whose identity and memory persist across sessions—so they can learn from experience over time rather than operating as stateless, single-session LLM chats. Letta presents its work as an agent runtime/harness for model-agnostic deployment: developers can connect LLMs of choice to a persistent agent system that manages context and long-lived memory. Letta’s flagship direction is Letta Code, which it describes as a model-agnostic agent harness with persistent memory, plus a developer SDK and tooling that support building and deploying agents that retain and evolve their state across model providers and machines. The company originally emerged from MemGPT research and positions itself as evolving beyond “memory as a plugin” toward memory and context management as core responsibilities of the agent harness itself. Letta’s recent product messaging emphasizes (1) a memory-first, inspectable architecture, (2) a shift from server-centric features toward runtime-native and client-side “computer use” workflows, and (3) extensibility via skills (and harness-level extension mechanisms such as Mods) so agent capabilities can be composed and improved over time. Letta targets developers and teams building long-running agent applications (including coding agents), and it also supports production deployment by offering persistent agents that can run across different model endpoints while preserving the agent’s memory and identity.

News

Company record

Aug 01, 2026 · Official · Letta BlogLetta Agents SDK: An SDK for stateful, persistent agents

Letta announced an SDK positioned for building stateful, persistent agents that keep identity and memory across models, machines, and interfaces.

Jun 01, 2026 · Official · Letta BlogIntroducing Mods: Enabling Agents to Self-Improve through Harness-Level Adaptation

Letta introduced Mods as an agent-friendly way to extend and adapt the Letta Code harness with harness-level adaptation to support agent self-improvement.

Apr 06, 2026 · Official · Letta BlogIntroducing the Letta Code App

Letta launched the Letta Code app as a way to interact with persistent coding agents that learn over time and work locally.

Mar 16, 2026 · Official · Letta BlogLetta's Next Phase

Letta outlined a strategic refocus around Letta Code as a flagship model-agnostic harness, describing how memory and workflows are evolving toward runtime-native approaches (including git-backed context repositories and client-side sleep/dreaming workflows).

Mar 01, 2026 · Official · Letta BlogOrchestrating Claude Code & Codex Agents

Letta published guidance on orchestrating coding agents with memory-first, persistent context, positioning it as a way to handle cases where stateless coding agents may differ.

Dec 16, 2025 · Official · Letta BlogLetta Code: A Memory-First Coding Agent

Letta released/announced Letta Code as a memory-first coding agent designed to persist across sessions and improve with use, and it described Letta Code as a model-agnostic OSS coding harness.

Oct 23, 2025 · Official · Letta BlogLetta Evals: Evaluating Agents That Learn

Letta introduced Letta Evals as an open-source evaluation framework intended to systematically test stateful agents that learn over time.

Sep 24, 2024 · Official · Letta BlogAnnouncing Letta

Letta’s public introduction post describing the framework for stateful agents and how it addresses the limitations of stateless LLM interactions.

Source map · 5 recurring channels · 15 references

Still resolving: Newsroom

What it builds

Letta Code

Letta Code is Letta’s memory-first, model-agnostic agent harness/runtime described as providing persistent memory plus capabilities such as computer use, skills, subagents, and transparent memory systems, with deployment paths across model providers.

source ↗
Letta Code app

The Letta Code app is a desktop application for interacting with deeply personalized, persistent Letta Code agents that learn over time and work locally on the user’s machine.

source ↗
Letta Agent SDK

The Letta Agents SDK is described as an SDK for stateful, persistent agents that keep identity and memory across models, machines, and interfaces.

source ↗
Mods

Mods are described as an agent-friendly way to extend and adapt the Letta Code harness at the harness level to enable agent self-improvement via harness-level adaptation.

source ↗
Conversations API

The Conversations API is described as allowing developers to build agents that can maintain shared memory across concurrent experiences with users.

source ↗
Letta Agent

Letta’s product section references Letta Agent as a self-improving AI agent whose memory, identity, and capabilities evolve with experience.

source ↗
Agent File (.af)

Agent File (.af) is a portable, shareable file format for representing agentic state (including tools, execution environment, model configuration, and memories) so agents can be recreated across servers.

source ↗
Letta Filesystem / context repositories (MemFS direction)

Letta’s product direction describes moving toward “git-backed memory” and a filesystem-like approach (context repositories) where memory is operated on through runtime-native tools rather than specialized memory tools.

source ↗

Milestones & partnerships

Apr 06, 2026
Launched Letta Code app

Letta launched the Letta Code desktop application for interacting with persistent coding agents locally and switching models while preserving agent memory and identity.

Mar 16, 2026
Published Letta’s Next Phase strategy update

Letta outlined a refocus around Letta Code as an open, model-agnostic harness and described planned deprecations and transitions of server-side features toward client-side and runtime-native replacements.

Dec 16, 2025
Positioned Letta Code as flagship memory-first coding agent

Letta introduced Letta Code as a memory-first coding agent designed to persist across sessions, and it stated Letta Code’s model-agnostic harness positioning on a coding benchmark.

Jul 01, 2025
Announced direction toward Letta Filesystem / context repositories

Letta described its shift toward using git-backed files/context repositories and a filesystem-like approach for memory management (MemFS direction) as part of its next-phase positioning.

Apr 02, 2025
Released Agent File (.af) for portable agent state

Letta launched Agent File (.af) as a file format intended to make stateful agents shareable and reproducible across servers.

Sep 24, 2024
Seed financing round formally announced

PR Newswire announcement of Letta’s $10M seed round led by Felicis with participation from Sunflower Capital and Essence VC, and named angel investors including Jeff Dean and Clem Delangue.

Felicis · Sunflower Capital · Essence VC
Sep 23, 2024
Exited stealth with seed funding announcement

Letta publicly emerged from stealth and announced a $10M seed round led by Felicis to build AI with advanced memory systems.

Felicis · Sunflower Capital · Essence VC