AI Frameworks · letta-ai

letta

Letta is an open-source Python platform for building stateful AI agents with persistent memory that can learn and improve over time. It supports local execution, cloud deployment, and integration into custom applications via SDKs, with model-agnostic architecture and pre-built skills.

Talk to DEV.co Discuss Implementation

Source: GitHub — github.com/letta-ai/letta

23.7k

GitHub stars

2.5k

Forks

Python

Primary language

Apache-2.0

License (OSI-approved)

Key facts

Objective fields from the source. Values we can't verify are shown as “Unknown” rather than guessed.

Field	Value
Repository	letta-ai/letta
Owner	letta-ai
Primary language	Python
License	Apache-2.0 — OSI-approved
Stars	23.7k
Forks	2.5k
Open issues	49
Latest release	0.16.8 (2026-05-14)
Last updated	2026-07-03
Source	https://github.com/letta-ai/letta

What letta is

Letta provides agent orchestration via a TypeScript Agent SDK and Python CLI, with configurable backends (local, cloud, self-hosted App Server). The repository contains the legacy V1 API server; active development has shifted to the separate letta-code repository. Agents support skills, subagents, and streaming interfaces with pluggable LLM support.

Quickstart

Get the letta source

Clone the repository and explore it locally.

terminalbash

1git clone https://github.com/letta-ai/letta.git2cd letta3# follow the project's README for install & configuration

Need it deployed, integrated, or customized instead? DEV.co ships production installs.

Best use cases

Stateful conversational AI with persistent memory

Build chatbots and assistants that retain context across sessions and improve through interaction history, ideal for customer support and personalized workflows.

Autonomous task agents with self-improvement

Deploy agents that can execute code tasks, manage subagents, and refine behavior over time—suitable for DevOps automation, coding assistance, and complex workflows.

Model-agnostic multi-tenant agent hosting

Leverage the Agent SDK to embed agents in SaaS or enterprise applications, supporting multiple LLM backends (Anthropic, OpenAI, etc.) with managed or self-hosted infrastructure.

Implementation considerations

Clarify whether your use case targets the legacy V1 API server (this repo) or the newer Letta Agent SDK and letta-code—the README recommends the latter for new projects.
Choose backend upfront: local execution requires Node.js 22.19+; cloud requires Letta API key; self-hosted App Server demands operational setup (container orchestration, storage, networking).
Understand LLM cost and latency implications—agents send context/history to external models. Monitor token usage and implement caching/pruning strategies for long-lived agents.
Plan for state management: persistent memory requires a backing store (database, file system, or managed service). Evaluate data durability, retention policies, and access controls.
Test agent behavior thoroughly—LLM outputs are non-deterministic and can drift. Implement monitoring, logging, and human-in-the-loop validation for production use.

When to avoid it — and what to weigh

You need a production-grade V1 API server without changes — The README explicitly states this repository is legacy; active development moved to letta-code. V1 API server maintenance/support status is unclear.
You require strict air-gapped or on-premise-only deployment — Default quickstart examples assume cloud API keys and Constellation backend. Self-hosting via App Server is possible but requires additional setup and operational overhead.
You need guaranteed compliance certifications or security audits — No security certifications, penetration testing results, or compliance audit reports are documented. Legal terms are external (privacy policy, ToS) without embedded compliance details.
You prioritize stability over bleeding-edge features — Repository shows active but rapid development (49 open issues, recent major updates). Early versions (0.16.x) may have breaking changes; backward compatibility guarantees are not stated.

License & commercial use

Licensed under Apache License 2.0 (Apache-2.0), a permissive OSI-approved open-source license allowing commercial use, modification, and distribution with attribution and liability disclaimers.

Apache 2.0 permits commercial use. However, this repository is noted as legacy; verify that your use case aligns with ongoing letta-code development. Consult the external terms of service (https://www.letta.com/terms-of-service) and privacy policy for any restrictions on Letta cloud services. Self-hosting the App Server may carry separate operational and compliance obligations.

DEV.co evaluation signals

Editorial assessment — not user reviews. Directional, with an explicit confidence level.

Signal	Assessment
Maintenance	Active
Documentation	Adequate
License clarity	Clear
Deployment complexity	Moderate
DEV.co fit	Good
Assessment confidence	High

Security considerations

No security audits, penetration testing, or threat model documentation are provided. Agent execution involves dynamic code/skill execution—input validation and sandboxing strategy are not detailed. External LLM calls expose context/memory to third-party APIs; ensure data classification and retention policies align with enterprise requirements. Self-hosted deployments inherit cloud provider and infrastructure security posture. Audit agent permissions, credential rotation (API keys), and audit logging before production deployment.

Alternatives to consider

LangChain / LangGraph

Well-established agent/chain framework with broader integration ecosystem, stronger docs, and larger community. Lacks built-in persistent memory but more flexible for custom state management.

AutoGen (Microsoft)

Multi-agent conversation and orchestration framework, strong enterprise backing, and mature feature set. Less focused on memory/learning but simpler deployment model.

AgentKit / Claude SDK

Lightweight, model-specific (Claude), minimal dependencies. Ideal if you're committed to Anthropic and prefer simplicity over advanced memory/multi-agent features.

Software development agency

Build on letta with DEV.co software developers

Letta offers a flexible, open-source foundation for stateful agents. Confirm that your use case aligns with the newer letta-code architecture (recommended for new projects), assess backend and data persistence requirements, and audit security/compliance needs before production deployment.

Talk to DEV.co

Related open-source tools

Surfaced by semantic similarity across the DEV.co open-source index.

openai-agents-python opensquilla EverOS

Related on DEV.co

Explore the category and the services that help you build with it.

AI Frameworks AI Application Development API Development Cloud Deployment Talk to DEV.co

letta FAQ

Should I use this repository (V1 server) or letta-code for new projects?

The README explicitly recommends letta-code and the Agent SDK for new projects. This repository is the legacy V1 API server; use only if maintaining existing V1 deployments or if you specifically need the older API surface.

Can I run Letta fully offline?

Yes, via local backend (Agent SDK or CLI). Requires Node.js 22.19+ and a local LLM or cached model. Cloud/Constellation deployments require internet and Letta API keys. Self-hosted App Server can be deployed on-premises but depends on infrastructure.

What is the cost to deploy on Constellation vs. self-host?

Not documented in the README. Constellation (cloud) likely has per-request or subscription pricing; self-hosted App Server costs are infrastructure-dependent. Check https://letta.com and docs for pricing.

How do I migrate from V1 SDK to Agent SDK?

Migration guide is not provided in the README. Refer to external documentation (docs.letta.com) and the Discord community for guidance; breaking changes are likely.

Software development & web development with DEV.co

Need help beyond evaluating letta? DEV.co is a software development agency offering software development services and web development for teams of every size. Our software developers and web developers build custom software, web applications, APIs, and ai frameworks integrations — and maintain them long-term.

Evaluate Letta for Your Agent Workflow

Talk to DEV.co Schedule a Consultation