A curated list of awesome Letta projects, tools, tutorials, and resources for building stateful AI agents with persistent memory.
Letta is the leading platform for building AI agents with persistent memory. Unlike traditional chatbots that forget everything between sessions, Letta agents remember, learn, and improve over time.
- Official Resources
- Tutorials & Guides
- Example Projects
- Tools & Integrations
- Research & Papers
- Videos & Talks
- Community
- Contributing
- Letta Website - Official website and product information
- Letta Documentation - Comprehensive documentation and API reference
- Letta GitHub - Official Letta repository
- Letta Cloud - Hosted Letta platform
- Letta Blog - Official blog with technical deep-dives and announcements
- Letta Leaderboard - Leaderboard to determine how to choose the best language models for your agent
- Letta Python SDK - Python SDK for the Letta API
- Letta TypeScript SDK - TypeScript SDK for the Letta API
- Letta Agents SDK - SDK for stateful agents with identity, memory, and experience that persist across models, machines, and interfaces [Blog post]
- Agent SDK v1 to v2 Migration Guide - Migration examples for moving Agent SDK v1 integrations to v2
- AI Memory SDK - A lightweight agent memory SDK for Letta for adding agentic memory and learning in a pluggable way
- Learning SDK - Drop-in SDK for adding continual learning and long-term memory to any LLM agent
- Letta Evals - Open-source evaluation framework for systematically testing stateful agents [Blog post]
- Letta Quickstart - Official quickstart guide for the Letta API
- Letta Code Quickstart - Get started with the Letta Code CLI and desktop app
- DeepLearning.AI Course - Free course on agent memory with Letta
- Create Letta demo applications - Create Letta App lets you create apps with Letta
- Memory Architecture Design - Guide to designing effective memory blocks
- Letta's Filesystem - Using Letta's filesystem capabilities
- Context Repositories - Git-based memory and programmatic context management in Letta Code
- Multi-Agent Systems - Building coordinated agent teams
- Custom Tools Development - Creating custom tools for agents
- Claude Subconscious - Give Claude Code a subconscious: a background memory and retrieval runtime for coding agents
- thought stream ATProto web viewer - Web viewer for the ATProto
stream.thought.blipchatroom experiment - thought stream ATProto CLI - Rust terminal client for reading and publishing
stream.thought.bliprecords in the ATProto chatroom experiment
- Letta OSS UI - An open-source demo UI built on top of the Letta Agents SDK
- Letta Agents SDK React Chat - A complete React chat example built with the Letta Agents SDK
- Agent SDK mobile app - Example mobile app built on top of the Letta Agents SDK, supporting both Letta Cloud and self-hosted servers
- Letta Chatbot template - An example Next.js chatbot application built on the Letta API, which makes each chatbot a stateful agent (agent with memory) under the hood (archived)
- Vercel AI SDK Provider Examples - AI SDK integration examples
- Discord bot - An example Discord chatbot built on the Letta API, which uses a stateful agent (agent with memory) under the hood
- CharacterPlus - Example CharacterAI-style web app that runs on Letta to create characters with memory
- Deep research agent - Open-source deep research agent implemented with Letta
- DuckDB agent - Example text-to-SQL agent using MotherDuck (DuckDB)
- Example social agent - An example stateful AI agent powered by Letta and Gemini 3 Pro
- Co - A thinking partner app built on Letta
- Signal Desktop Letta fork - A fork of Signal Desktop with Letta agents embedded in Signal conversations
- Context Constitution - A set of principles governing how AI agents manage context to learn from experience [Blog post]
- Letta Code - Memory-first, model-agnostic open source coding agent with persistent memory, skills, subagents, and mods [Blog post]
- Letta Code Desktop app - Desktop app for interacting with deeply personalized agents that learn over time
- Skills - A shared repository of skills for teaching agents new tasks, usable with Letta Code and other harnesses
- Mods - Mod packages and examples for Letta Code: agent-created extensions that adapt the harness itself [Blog post]
- Letta ACP - ACP (Agent Client Protocol) adapter for using stateful Letta agents from Zed and other ACP editors
- Letta Code GitHub Action - Integrate Letta Code into your GitHub repo to review issues, code, and more
- Trajectory - Convert sessions across harnesses to a unified trajectory format
- Remote Environments - Message an agent working on your laptop from your phone
- Vercel AI SDK Provider - Use Letta with Vercel AI SDK v5
- Zapier Integration - Use Letta with Zapier
- Telegram - A Modal application for serving a Letta agent on Telegram
- Obsidian - An Obsidian plugin for serving a Letta agent on Obsidian
- n8n - Connect your Letta agent to n8n workflows
- Letta Voice - Chat with your Letta agents over a low-latency voice connection
- Flows - Experimental TypeScript orchestration DSL for the Letta Agent SDK, with conversation branching and forking, bounded parallelism, and structured results
- Social CLI - A unified CLI to connect artificial intelligence to the social web
- Hypervigilant - A file watcher that sends saved diffs to persistent Letta agents
- Your tool here! - Submit a PR
- Letta Computer Deployment - Deploy an always-on computer connected to Letta Cloud
- Letta App Server Deploy - Deploy a directly accessible Letta App Server to any cloud platform
- Letta Code CLI - Command-line interface for Letta agents
- Agent Development Environment - Web-based agent IDE
- MemGPT: Towards LLMs as Operating Systems - Original MemGPT paper
- Sleep-time Compute - Research on agent sleep-time compute, with accompanying GitHub repository and blog post
- Memory Models: Towards Agents That Learn - Powering agent learning through memory models trained with memory-native RL
- Continual Learning in Token Space - Learning in token space as the key to building AI agents that truly improve over time
- Context Repositories: Git-based Memory - A rebuild of agent memory using programmatic context management and git-based versioning
- Context Constitution - Principles governing how AI agents manage context to learn from experience
- Terminal-Bench - Letta integration for terminal-bench [Blog post]
- Recovery-Bench - A benchmark for evaluating the capability of LLM agents to recover from mistakes [Blog post]
- Memory Models: Towards Agents That Learn - June 25, 2026
- Why Memory Isn't a Plugin - April 2026
- Context Repositories: Git-based Memory for Coding Agents - February 12, 2026
- Continual Learning in Token Space - December 2025
- Rearchitecting Letta's Agent Loop: Lessons from ReAct, MemGPT, & Claude Code - October 14, 2025
- Introducing Recovery-Bench: Evaluating LLMs' Ability to Recover from Mistakes - August 27, 2025
- Benchmarking AI Agent Memory: Is a Filesystem All You Need? - August 12, 2025
- Building the #1 Open Source Terminal-Use Agent Using Letta - August 5, 2025
- Agent Memory: How to Build Agents that Learn and Remember - July 7, 2025
- Anatomy of a Context Window: A Guide to Context Engineering - July 3, 2025
- Letta Leaderboard: Benchmarking LLMs on Agentic Memory - May 29, 2025
- Memory Blocks: The Key to Agentic Context Management - May 14, 2025
- Sleep-time Compute - April 21, 2025
- RAG is not Agent Memory - February 13, 2025
- Stateful Agents: The Missing Link in LLM Intelligence - February 6, 2025
- The AI Agents Stack - November 14, 2024
- Letta Agents SDK: An SDK for Stateful Agents - August 17, 2026
- Introducing Mods: Enabling Agents to Self-Improve through Harness-Level Adaptation - June 24, 2026
- Introducing the Letta Code App - April 6, 2026
- Building Draft: A Personal Assistant that Learns - March 2026
- Orchestrating Claude Code & Codex Agents - March 16, 2026
- Letta's Next Phase - March 16, 2026
- Remote Environments for Letta Code - March 4, 2026
- Conversations: Shared Agent Memory Across Concurrent Experiences - January 2026
- Letta Code: A Memory-First Coding Agent - December 2025
- Programmatic Tool Calling with Any LLM - December 1, 2025
- Letta Evals: Evaluating Agents That Learn - October 2025
- Introducing Letta Filesystem - July 24, 2025
- Announcing Letta Client SDKs for Python and TypeScript - April 17, 2025
- Introducing the Agent Development Environment - January 15, 2025
- Announcing Letta - September 23, 2024
- MemGPT is now part of Letta - September 23, 2024
- The Letta YouTube channel - Official Letta YouTube channel
- Letta Memory Tool Demo - Agents that redesign their own architecture
- How to use Archival Memory - How to use Letta's archival memory for storing long-term memory
- Building a self-improving deep research agent - How to build a self-improving deep research agent using Letta in just a few minutes
- An introduction to personality design - How to design a personality for your Letta agent
- The basics of memory architecture - How to iteratively improve a Letta agent's memory architecture
- Adding knowledge graphs with neo4j to Letta - Use MCP and Letta Desktop to build a knowledge graph
- How to use the Zapier integration - Connect your Letta agent to any external service
Weekly recorded office hours covering the latest Letta features and community Q&A:
- Office Hours: Free Dreaming, ACP, MCP, and Trajectory - July 2026
- Office Hours: Mods, MemFS, and Teaching Agents New Tricks - June 2026
- Office Hours: Shared Memory, Hypervigilant, Schedules, and the Agent SDK - 2026
- Stateful Agents Meetup: Networks - Recording of the Stateful Agents Meetup hosted by Letta and Nokia
- Building with Letta Agents - Community-maintained, AI-assisted explanatory guides to persistent agents and the Letta ecosystem; independent of official documentation
- Office Hours Guides - Community-maintained, AI-assisted guides to Letta Office Hours, linking explanations to the original recordings
- Discord - Official Discord community
- Forum - Official Forum
- Bluesky - Official Bluesky profile
- Twitter/X - Follow for updates
- Community Showcase - #showcase channel in Discord
- Interact with your Letta agent using n8n and Telegram - Guide to connect Letta agents with n8n and Telegram
- Submit your projects here via PR!
Contributions are welcome!
How to contribute:
- Fork this repository
- Add your resource in the appropriate category
- Ensure your addition follows the format:
[Resource Name](url) - Brief description - Submit a pull request
Criteria for inclusion:
- Must be related to Letta or stateful agent development
- Must be functional and actively maintained (for tools/projects)
- Must provide value to the Letta community
- Preferably open source (for projects and tools)