Letta
Click to visit website
About
Letta is an AI memory platform and operating system designed to solve the persistence problem inherent in standard large language models. While typical LLMs are stateless and lose information once a session concludes, Letta provides a structured framework for creating stateful agents that maintain context across interactions. Founded by the researchers behind MemGPT at UC Berkeley's Sky Computing Lab, the platform acts as a sophisticated middle layer that manages an agent's long-term memory, allowing it to evolve, learn from historical data, and retain a consistent personality or knowledge base over time. The system operates on a model-agnostic framework, meaning developers can connect various LLMs while keeping the agent's underlying memory intact. A core component of the ecosystem is Letta Code, a memory-first coding harness that allows agents to interact with local filesystems and learn from previous development tasks. The platform also introduces Context Repositories, which utilize git-based versioning to manage programmatic context dynamically. Additionally, Letta incorporates research-backed concepts like sleep-time compute, which allows agents to refine their memories and precompute potential responses during idle periods, effectively mimicking human cognitive cycles. This tool is primarily built for developers, AI engineers, and researchers moving beyond basic chatbots to build production-grade agents. It is particularly effective for creating digital humans, persistent coding assistants, and automated customer support systems where long-term relationship building and context retention are vital. Because it supports both local execution via a CLI and cloud-based management through a low-level API, it caters to solo hobbyists looking for personalized agents as well as enterprise teams requiring scalable, stateful AI infrastructure with high reliability. What sets Letta apart from traditional RAG (Retrieval-Augmented Generation) systems is its treatment of memory as a dynamic, self-improving state rather than a static document store. It transitions AI from a transactional tool into a perpetual entity. With its strong open-source foundation and model-agnostic nature, Letta ensures that developers are not locked into a specific provider's ecosystem, allowing an agent's history and intelligence to persist even if the underlying model is upgraded or changed.
Pros & Cons
Supports persistent memory that survives session resets and model changes.
Backed by an active open-source community with over 21,000 GitHub stars.
Offers a free tier with Bring Your Own Key (BYOK) support for local execution.
Includes innovative sleep-time compute features for offline agent learning and refinement.
The model-agnostic approach prevents vendor lock-in for your agent's state and history.
The Free plan is restricted to only three stateful agents at a time.
Server-side tool execution on the API incurs credit costs based on CPU usage.
The Max plan is strictly limited to personal use and not for building external applications.
Enterprise-grade security features like RBAC and SSO are missing from the lower tiers.
Use Cases
Software developers can use Letta Code to build persistent coding assistants that learn from their specific codebase over time.
AI researchers can utilize the framework to experiment with continual learning and memory management in a model-agnostic environment.
Customer support teams can deploy digital humans that maintain context and relationship history across multiple user interactions.
Solo creators can develop highly personalized agents that run locally on their filesystem with customized long-term memory.
Enterprise organizations can leverage the API and stateful agents to create internal knowledge assistants with secure SSO integration.
Platform
Features
• model-agnostic architecture
• low-level memory management api
• local filesystem access
• perpetual learning framework
• letta code cli
• sleep-time compute refinement
• git-based context repositories
• stateful agent memory
FAQs
How is tool execution charged on the Letta API?
Server-side tools executed on the Letta API incur credit costs based on the CPU time required for execution. However, remote MCP tools, built-in tools, and client-side tools run via Letta Code do not incur any credit costs.
Can I use my own LLM with Letta?
Yes, Letta is model-agnostic and specifically allows users on the Free plan to connect their own LLM API keys through a Bring Your Own Key (BYOK) system.
What is the difference between the Pro and Max plans?
The Pro plan is for building personal agents with basic monthly credits, while the Max plan is designed for power users running token-intensive workloads like high-throughput coding with significantly higher request limits.
What exactly is Letta Code?
Letta Code is a memory-first coding harness built on the Letta API that allows agents to learn over time and access local filesystems via a CLI, making them portable across different models.
Does Letta offer any security features for organizations?
Yes, the Enterprise plan includes professional security and management features such as Role-Based Access Control (RBAC), SSO via SAML/OIDC, and dedicated support for high-volume needs.
Pricing Plans
Pro
USD20.00 / per month• Unlimited stateful agents
• Access frontier models on Letta API
• $20 of monthly API credits
• Pay-as-you-go credit overage
Max
USD200.00 / per month• Higher usage limits for power users
• Optimized for Letta Code
• Early access to new features
• Request quotas reset multiple times daily
Enterprise
Unknown Price• Volume-based pricing
• Increased quotas
• Role-based access control (RBAC)
• SAML/OIDC SSO authentication
• Dedicated support
Free
Free Plan• 3 stateful agents
• Connect own LLM API keys (BYOK)
• Chat with agents in the ADE
• Run agents locally with Letta Code
Job Opportunities
Product Engineer, Agent Development Environment
Enable AI agents to remember every interaction and learn continuously with a memory-first operating system designed for developers building perpetual intelligence.
Benefits:
Paid in-person work trial
Experience Requirements:
Strong frontend development skills (React)
Experience with fullstack development including API design
Proficiency with backend development (Next.js, Node, Typescript)
Interacting with databases (Postgres, Redis)
Experience crafting user experiences
Other Requirements:
Excellent communication skills
Excitement about AI agents
Ability to work in-person 5 days a week in San Francisco
Interest in how agents and LLMs work under the hood
Responsibilities:
Build, design and define how developers interface with agent framework
Build visual interfaces for innovative agent functionality
Interface with customers and open source community
Develop application SDKs for full-stack agent apps
Develop Cloud Platform for deployment and monitoring
Show more details
Software Engineer, Agent Platform
Enable AI agents to remember every interaction and learn continuously with a memory-first operating system designed for developers building perpetual intelligence.
Benefits:
Paid in-person work trial
Experience Requirements:
Strong proficiency with Python
Understanding of architecting services for security, reliability, and performance
Ability to build scaleable backend services
Strong understanding of SQL databases (Postgres)
Familiarity with AI stack tooling (vLLM, Ollama, OpenAI APIs)
Other Requirements:
Ability to work in-person 5 days a week in San Francisco
Proficiency with TypeScript and React is a bonus
Responsibilities:
Develop Letta's OSS agents framework and cloud service
Design scaleable & resilient backend services and APIs
Work with researchers to support bleeding-edge agent architectures
Define developer APIs for agents
Lead development of company's OSS stack
Show more details
Research Engineer / Scientist, Post-Training
Enable AI agents to remember every interaction and learn continuously with a memory-first operating system designed for developers building perpetual intelligence.
Benefits:
Paid in-person work trial
Experience Requirements:
Proficiency in Python and deep learning frameworks (PyTorch)
Expertise in post-training techniques (SFT, RL, reward models)
Proven track record of impactful research
Real-world impact beyond pure academic work
Other Requirements:
Ability to balance execution speed with empirical rigor
Ability to work in-person 5 days a week in San Francisco
Responsibilities:
Training models for better agentic tool-use
Designing mechanisms for continuous model weight updates
Designing and running experiments for data mixtures and models
Building infrastructure for generating synthetic data at scale
Building challenging evals for measuring agentic capabilities
Show more details
Ratings & Reviews
No ratings available yet. Be the first to rate this tool!
Alternatives
Pieces
Automatically capture and recall every coding snippet, research link, and chat message with a local-first AI memory engine designed to boost developer flow.
View DetailsFeatured Tools
adly.news
Connect with engaged niche audiences or monetize your subscriber base through an automated marketplace featuring verified metrics and secure Stripe payments.
View Details
Atoms
Launch full-stack products and acquire customers in minutes using a coordinated team of AI agents that handle everything from deep research to SEO and coding.
View Details
GenMix
Generate professional-quality AI videos, images, and voiceovers using world-class models like Sora 2 and Kling 2.6 through a single, unified creative dashboard.
View Details
Reztune
Land more interviews by instantly tailoring your resume to any job description using AI-driven keyword optimization and professional, ATS-friendly templates.
View Details
Image to Image AI
Transform photos and videos using advanced AI models for face swapping, restoration, and style transfer. Perfect for creators needing fast, professional visuals.
View Details
Nano Banana
Edit and enhance photos using natural language prompts while maintaining character consistency and scene structure for professional marketing and digital art.
View Details
Nana Banana Pro
Maintain perfect character consistency across diverse scenes and styles with advanced AI-powered image editing for creators, marketers, and storytellers.
View Details
Kling 4.0
Transform text and images into cinematic 1080p videos with multi-shot storytelling, character consistency, and native lip-synced audio for professional creators.
View Details
AI Seedance
Generate 15-second cinematic 2K videos with physics-based audio and multi-shot narratives from text or images. Ideal for creators and marketing teams.
View Details
Mistrezz.AI
Engage in immersive NSFW roleplay and ASMR voice sessions with adaptive AI companions designed for structured escalation, fantasy scenarios, and personal connection.
View Details
Seedance 3.0
Transform text prompts or static images into professional 1080p cinematic videos. Perfect for creators and marketers seeking high-quality, physics-aware AI motion.
View Details
Seedance 3.0
Transform text descriptions into cinematic 4K videos instantly with ByteDance's advanced AI, offering professional-grade visuals for creators and marketing teams.
View Details