PC Slower Than It Used to Be?
A free scan shows the junk files, broken settings and background clutter dragging Windows down - then fixes them in one click.Free scan · Windows 10 & 11Crashes, No Sound, or Screen Glitches?
Random freezes, missing sound and display glitches usually trace back to one bad driver. Find and replace yours safely.Free scan · under a minuteLetta emerged from stealth on September 23, 2024, with a $10 million seed round led by Felicis at a reported $70 million post-money valuation. Founded by UC Berkeley researchers Sarah Wooders and Charles Packer, the company commercialized the open-source MemGPT project’s approach to persistent, editable memory for AI agents. By 2026, Letta’s emphasis had moved beyond hosted memory APIs toward Letta Code, a model-agnostic agent harness for local and remote computer work.
What Letta announced in September 2024
Letta’s launch was a financing and product announcement, not the debut of an ordinary chatbot. The company said it had raised $10 million in seed financing led by Felicis, with participation from Sunflower Capital and Essence VC. TechCrunch reported a $70 million post-money valuation. The financing release listed angel backers including Jeff Dean, Clem Delangue, Cristóbal Valenzuela, Jordan Tigani, Robert Nishihara, Tristan Handy and Barry McCardel.
| Launch fact | Verified detail |
|---|---|
| Emergence from stealth | September 23, 2024 |
| Seed round | $10 million |
| Lead investor | Felicis |
| Reported valuation | $70 million post-money |
| Founders | Sarah Wooders and Charles Packer |
| Research origin | UC Berkeley Sky Computing Lab |
| Commercial focus at launch | Stateful agents, developer tooling and Letta Cloud |
Sources: Letta’s announcement, the financing release and TechCrunch’s launch coverage.
Why the Berkeley connection mattered
Letta grew out of UC Berkeley’s Sky Computing Lab, led by Ion Stoica. TechCrunch described that research ecosystem as a successor to Berkeley’s RISELab and AMPLab, environments associated with projects and companies such as Anyscale, Databricks, SiFive, vLLM, Gorilla and SGLang.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
#1 Best Overall
- High-Performance AI Processor：The MS-02 Ultra features an Intel Core Ultra 9 285HX (24C/24T, up to 5.5 GHz, 13 TOPS NPU), delivering fast and efficient performance for AI inference, algorithm development, and media workloads. A PCIe x16 expansion slot supports desktop-class GPU upgrades for advanced model training and accelerated computing tasks. It's ideal for creators, engineers, and teams handling intensive parallel workloads.
- 4 × M.2 PCIe 4.0 + 4 × DDR5 SODIMM slots:Four DDR5 SODIMM slots support up to 256 GB of memory, while ECC helps maintain data integrity in mission-critical environments. Four PCIe 4.0 M.2 slots support up to 24 TB of storage, supporting RAID 0/1/5/10, combining high-speed performance with data protection. It allows for the creation of independent scratch disks, media libraries, and project drives, providing high-throughput for production workflows.
- PCIe & USB 4.0 v2: Up to three PCIe slots can be equipped, including a dual-slot x16 GPU. The main slot supports PCIe 5.0, meeting the needs of high-bandwidth creative and computing workloads. USB 4.0 v2 (80Gbps) supports high-bandwidth external storage and displays.
- Ultra-fast Networking: Wi-Fi 7 further enhances wireless performance with next-generation speeds and low-latency stability. Intelligent bandwidth switching optimizes throughput in different network environments, ensuring optimal performance for enterprise or local networks. Dual 25GbE ports (providing up to approximately 3.125 GB/s bandwidth, about 25 times faster than traditional 1GbE), enabling seamless large-scale file transfers and parallel computing. 10GbE and 2.5GbE ports, with support for Intel vPro technology, ensure enterprise-grade remote management and deployment flexibility.
- Server-grade thermal architecture: Utilizing a dedicated CPU/GPU airflow design, equipped with a 6-pipe dual-fan cooler, it maintains stable performance even under sustained loads, delivering up to 140W Turbo power while maintaining a 100W TDP, and operating with noise levels as low as 36 dB. An integrated 350W power supply ensures stable and reliable output for demanding computing tasks and fully loaded extended configurations.
That history helped Letta attract attention because the lab connected academic systems research with widely used open-source infrastructure. It does not mean UC Berkeley owned Letta. The founders and underlying research were Berkeley-connected, while the company commercialized the work independently. The lab’s republication of the launch article confirms the relationship without establishing a university corporate spinout process.
Source: UC Berkeley Sky Computing Lab.
MemGPT was the research project behind the company
MemGPT began as a Berkeley research project asking whether a language model could manage information beyond its immediate context window. Its design used tools and an operating-system-like memory hierarchy: information could be moved between a limited working context and longer-lived storage under the agent’s control.
The project became widely known after its 2023 paper and open-source code. Letta later explained that MemGPT was the research name and design pattern, while Letta became the company and framework name. The Python package moved to letta, and the Docker image to letta/letta-server; the open-source repository continued to be maintained.
Source: Letta’s MemGPT-to-Letta explanation.
The problem Letta was trying to solve
Most language-model APIs are stateless from an application’s point of view. A developer must repeatedly provide conversation history, user information, documents and intermediate results. That works for a single exchange, but becomes costly and fragile when an agent must operate over weeks, personalize its behavior or coordinate a long-running task.
Free tools Windows power users keep installed
One-click scans. No signup required.
Letta’s proposed architecture made several pieces explicit:
- Persistent memory: durable memory blocks or other state that survive individual requests.
- Agent-controlled updates: the agent can decide what to retain, revise, retrieve or discard.
- Tools and external data: the agent can call functions and data sources rather than relying only on text in a prompt.
- Stateful APIs: an identified agent can continue across interactions.
- Provider separation: application state can remain while the underlying model changes.
- Visibility: developers can inspect and edit context instead of treating memory as an opaque vendor feature.
This is application-managed memory, not human-like memory. It can preserve stale information, store an inference as fact, expose private data or compound an earlier mistake.
Persistent memory is different from a longer context window
| Concept | What it does |
|---|---|
| Long context | Places more tokens in the model’s current input. |
| Persistent memory | Stores selected information across sessions and retrieves or rewrites it later. |
| Agent state | Combines memory with identity, tools, tasks and execution history. |
| Model independence | Allows state to persist while the application changes model providers. |
Letta’s thesis was that increasing a context window does not by itself provide a controllable system for deciding what should survive, how it should be edited or how it should be shared between agents. Separating data and state from model computation also creates a potential infrastructure layer independent of any one foundation-model vendor.
What Letta offered at launch
The September 2024 product plan combined open-source software with a hosted service:
- An open-source framework derived from MemGPT.
- An Agent Development Environment for building, debugging and deploying stateful agents.
- Letta Cloud for hosted agent execution and persistent state.
- REST APIs for stateful agents.
- Connections to external inference providers, including OpenAI, Anthropic and vLLM.
Letta Cloud was still being piloted and accepting beta-user requests when the company announced its financing; it was not presented as a fully open, generally available business service on launch day. Source: Letta’s launch announcement and TechCrunch.
How the strategy compared with other agent stacks
OpenAI’s integrated tools
OpenAI’s agent products can be simpler for organizations already standardized on OpenAI models and infrastructure. Letta’s launch positioning emphasized a more independent state and memory layer that could use multiple providers. That is a strategic contrast, not proof that one approach is universally better.
Rank #3
- Professional AI & Creator Workstation: AMD Radeon AI PRO R9700 GPU with 32GB GDDR6 is engineered for AI development, professional content creation, and compute-intensive workloads.
- Massive 32GB Memory Capacity: 32GB of GDDR6 memory on a 256-bit bus provides ample bandwidth for large AI models, 8K video editing, and complex 3D rendering.
- Advanced RDNA 4 with AI Accelerators: 64 Compute Units with 3rd Gen Ray Tracing and dedicated 2nd Gen AI Accelerators for groundbreaking AI performance and visual computing.
- Professional Blower Cooling: Efficient single blower design exhausts heat directly out of the chassis, ideal for multi-GPU workstation and server configurations.
- Enterprise-Grade Thermal Solution: Vapor chamber heatsink with industrial Honeywell PTM7950 thermal interface material ensures reliable cooling under sustained professional loads.
LangChain and LangGraph
LangChain and LangGraph provide a broad orchestration, workflow and integration ecosystem. They may be the better fit when graph-based control and integrations matter more than Letta’s memory-first design. Letta’s distinction is the emphasis on long-lived, inspectable agent state.
Provider-specific coding agents
Claude Code, Codex CLI and Gemini CLI can offer tighter integration with their native models. Letta Code’s counterargument is portability: the agent’s identity and memory can persist while the model changes. Portability does not guarantee identical tool use, latency, cost or output quality after a model switch.
Recommended Free Tools
What changed by 2026
Letta’s March 2026 strategy announcement described a new phase centered on Letta Code and a broader goal of building “machines that learn.” The product is presented as an open, model-agnostic agent harness with memory, computer use, skills, subagents and deployment. Letta also said some legacy features, including tool rules and legacy tools, were being deprecated, with further template and filesystem changes planned for April 2026.
The April 2026 Letta Code app announcement described local agents that work with files and projects, initialize memory from codebases and prior sessions, and use background memory subagents to review and refine context. Users can run /init to initialize memory and /doctor to reorganize or clean it. The app supports macOS, Windows and Linux, and can use a user’s own API keys or eligible coding plans from Codex/ChatGPT and Z.ai.
Sources: Letta’s next-phase announcement and Letta Code app announcement.
Rank #4
- FAST RUNS IN THE FAMILY — The 16-inch MacBook Pro with the M5 Pro or M5 Max chip brings next-generation speed and powerful on-device AI to personal, professional, and creative tasks. With all-day battery life, double the starting storage,* and a breathtaking Liquid Retina XDR display, it’s pro in every way.*
- BUCKLE UP — Along with a next-generation CPU, faster unified memory, and up to 2x faster SSD storage,* M5 Pro and M5 Max feature a more powerful GPU with a Neural Accelerator built into each core, delivering faster AI performance and on-device training capabilities. So you can blaze through demanding workloads at mind-bending speeds.
- BUILT FOR AI — Apple silicon, and every major component that powers it, is designed to run demanding on-device AI workloads like LLM inference and training. And Apple Intelligence helps you write, express yourself, and get things done effortlessly with groundbreaking privacy protections at every step.*
- ALL-DAY BATTERY LIFE — MacBook Pro delivers the same exceptional performance whether it’s running on battery or plugged in.*
- MACOS RUNS APPS FAST — All your go-to apps run lightning fast in macOS, including built-in apps like FaceTime and Messages. Plus, built-in virus protection and free software updates help keep your Mac running smoothly and securely.
What you can use today
Letta’s documentation presents three broad routes:
Do these 3 things before closing this tab:
1Repair Windows errors before they cause bigger problems2Scan for outdated or missing drivers - takes under a minute3Clear out junk files and repair common Windows errors- Letta Code: a local, memory-first coding and computer-use agent.
- Letta Agent SDK/API: tools for embedding stateful agents in an application.
- Letta Cloud or a self-hosted Letta App Server: deployment choices for running agents.
Install Letta Code
The documented CLI requires Node.js 18 or newer:
npm install -g @letta-ai/letta-code
lettaInside a project, run:
cd your-project
lettaThen use /init to bootstrap project memory. Other documented commands include /remember, /memory, /model, /search, /clear, /new, /resume and /agents. letta server can run the process on a remote machine accessed through chat.letta.com or the desktop app. Source: Letta’s desktop guide.
Use the SDK
pip install letta-clientnpm install @letta-ai/letta-clientHosted API requests require an API key and bearer authentication; the SDKs handle the request details. Sources: Python SDK documentation, TypeScript SDK documentation and API documentation.
Who is Letta a good fit for?
- Developers building agents that must persist across sessions.
- Teams needing inspectable, editable user or project memory.
- Organizations that want to change model providers without discarding application state.
- Users who need local or remote computer use, skills, subagents or scheduled work.
- Teams prepared to self-host when data control matters more than turnkey operations.
A conventional chatbot user may not benefit from the additional state and operational complexity.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Risks, costs and operational trade-offs
Memory can be wrong
Agents can retain obsolete project details, incorrect preferences or guesses presented as facts. Teams need policies for inspection, correction, deletion, conflict resolution and tenant isolation.
Best Value
- 【High-Performance APU】The MS-S1 MAX features an AMD Ryzen AI Max+ 395 APU, integrating a Zen 5 architecture CPU (up to 5.1GHz, 16C/32T, 64M L3 Cache), an RDNA 3.5 GPU, and an NPU (50 TOPS). The total system output is 126 TOPS. It provides powerful parallel computing capabilities for demanding AI workflows. It is ideal for running local LLMs, multimodal models, and computationally intensive tasks
- 【128GB UMA Memory】Equipped with up to 128GB of LPDDR5x-8000MT/s unified memory, it enables the CPU and GPU to access a shared, high-bandwidth memory pool with extremely low latency. Ideal for large-scale AI inference, 3D workloads, and complex timelines in video editing. It eliminates traditional VRAM bottlenecks, ensuring smoother data transfer during high-intensity computations. The UMA design maximizes performance stability under high loads
- 【Flexible Expansion】The MS-S1 MAX features USB4 V2 (up to 80Gbps), dual 10GbE LAN, HDMI 2.1 (up to 8K60), a full-length PCIe x16 expansion slot, and dual M.2 slots supporting up to 16TB RAID 0/1. Wi-Fi 7 provides stronger signal coverage and a more stable wireless experience. The slide-out design facilitates upgrades and maintenance. It easily adapts to personal, studio, or rack-mount enterprise environments
- 【High-Efficiency Cooling System】Utilizing an aerospace-grade aluminum alloy chassis, copper base plate, six heat pipes, dual turbine fans, and advanced PCM thermal conductive material, it maintains stable cooling performance even under continuous load. This system supports 130W continuous power and 160W peak power operation, with a built-in 320W power supply. It boasts multiple global certifications including CCC, FCC, UL, CE, and UKCA, ensuring stable and reliable operation in various environments
- 【Cluster Design】Two MS-S1 MAX units can be configured as a dual-unit cluster to run a large 235B Q4 model locally, achieving an output speed of 10.87 tok/s. Supporting 2U rack deployment, multiple MS-S1 MAX units can be cascaded into a distributed cluster to create a high-efficiency AI computing center. A cluster of four MS-S1 MAX units successfully ran a DeepSeek-R1 671B Q4 large model. A reserved cluster power-on interface allows for unified start-up and shutdown
Portability is practical, not magical
Changing models can alter instruction following, tool-calling behavior, context interpretation, latency, cost and output quality even when memory remains intact.
Computer access increases the blast radius
Local files, shells, browsers and remote machines make an agent more useful but increase exposure to prompt injection, destructive commands, credential leakage and accidental edits. Start in a controlled repository or sandbox and restrict permissions until the workflow is understood.
Hosted and self-hosted have different burdens
Hosted deployment reduces infrastructure work but introduces vendor, data-residency, uptime and pricing dependencies. Self-hosting offers more control while shifting upgrades, monitoring, security and scaling to the buyer.
Letta’s reviewed pages do not establish a universal public subscription price for Letta Code or a complete current Letta Cloud pricing table. Letta says Code can be tried with users’ own API keys or eligible coding plans; model usage, hosted deployment and storage may still create separate costs. Confirm quotas, retention, regions, support and enterprise terms before committing.
Bottom line
Letta’s importance lies in making memory and state first-class parts of the agent stack. Its 2024 launch turned Berkeley’s MemGPT research into a funded company with an open-source framework, hosted-agent ambitions and a model-provider-independent thesis. By 2026, Letta Code broadened that idea into persistent coding and computer-use agents. The technology is a meaningful alternative for developers who need durable, inspectable state—but it is not proof that reliable continual learning, privacy or autonomous software operation has been solved.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
