LangChain

🚀 Deep Agents: The Weekly Roundup 🚀 Dive into our new resources to help you build Deep Agents capable of handling complex, long-running tasks. ✏️ Context engineering is key to reliable agents. Deep agents need detailed context and prompts, and filesystems can help manage that context. We wrote up some strategies for using filesystems to improve agent reliability. Read the blog: https://lnkd.in/g2E66mzR 🤷♀️ What are Deep Agents? - We break down the key things to know when you’re building an agent to handle more complex tasks. Watch the video: https://lnkd.in/g6GdWUMC ⚽ Using Skills with Deep Agents CLI - Agent skills are now available in the Deep Agents CLI, enabling you to use the large and growing collection of public skills with your agents. Watch the video: https://lnkd.in/dKxT4Tqh & blog: https://lnkd.in/dMUHA2hf 📚 LangChain Academy Course - If you’re not sure how to get started, our free LangChain Academy course covers the four features that set Deep Agents apart: planning, file systems, sub-agents, and prompting. By the end, you'll design, implement, and deploy your own. Enroll now: https://lnkd.in/gVhdxn8W

⭐ Build Deep Agents in our LangChain Academy course ⭐ Many agents today follow the same simple pattern: run in a loop, call tools. That architecture works well enough, but it breaks down as tasks get more complex. Today, companies of all sizes – from startups to large enterprises – are building their own Deep Agents. These agents dive deeper. They’re able to plan complex tasks and carry them out over longer time horizons. There are four key features that set Deep Agents apart from regular agents: 1. Planning – keeps agents on track 2. File system – allows agents to offload context 3. Sub-agents – act as focused specialists 4. Prompting – provides agents with detailed instructions Our LangChain Academy course, Deep Agents, shows you how to combine these pieces. By the end, you’ll have the skills to design, implement, and deploy your own Deep Agent. Enroll for free ➡️ https://lnkd.in/gBnuRdSW

✨ 🏖️ New Sandboxes for Deep Agents let you safely execute arbitrary code and bash commands in a remote environment. They're useful when you need a clean, reproducible environment, want to run parallel or long-running tasks, or need to keep your local machine safe from untrusted code. In this video you’ll learn how to: 1. Set up a remote sandbox 2. Make a tool call to execute code 3. Send a response back to your Deep Agent https://lnkd.in/gKU5NQCy

What’s the difference between agent runtime, a framework, and a harness? In this post we expand on the patterns we’re seeing, when to use different approaches, and how things are evolving. Our mental model in brief: - Frameworks: abstractions that help build agents with LLMs (LangChain) - Runtime: infrastructure level with support for durable execution, streaming, human-in-the-loop and more (LangGraph) - Harness: general purpose agent (Deep Agents) Read the post: https://lnkd.in/gUwSr_gv

Getting AI agents to production can be messy. Rippling's Head of AI breaks down the reality: why deterministic workflows fall short, how to handle edge cases, and the feedback loops that actually matter—including how they use LangSmith to debug and iterate. Watch Harrison Chase and Ankur Bhatt talk through context engineering, deep agents, and what it takes to ship reliable AI agents at scale → https://lnkd.in/eGrhqgcy

🌟 Introducing the LangChain Experts Program — and our founding cohort! This new program brings together community members who go above and beyond to help others build with LangChain. Huge shout-out to our founding Experts: Dmitry Labazkin, Benjamin McHone, Paweł Twardziak, and Mohammad Mohtashim Khan — thank you for leading the way! Experts help keep our community thriving by answering questions in the Forum and joining conversations across Slack and LangChain Academy. Meet our Experts and explore their profiles: langchain.com/community

Meet Harrison Chase at re:Invent! He'll be at booth #524 on Wednesday, Dec 3rd (12-1pm) to talk all things LangChain. Whether you're already building with us or just getting started, come by and let's connect. See you at AWS re:Invent—book time with our team, schedule a demo, or just stop by: https://lnkd.in/eCv-87FA

💬 Join us on December 15th (or December 16th for some APAC time zones) for a collaborative Community Jam Session! We’ll be gathering your thoughts on what’s working well, and what isn’t, in langChain 1.0 and 1.1. 🔗 🦜 We’ll cover: - langchain 1.0: Now that it’s been out for a while and you’ve had time to explore, we’d love to hear your real-world feedback. - langchain 1.1: We’ll walk through the new features introduced in 1.1 and gather your thoughts if you’ve tried them. Your feedback shapes the future of langchain — see you there! 👉 RSVP: https://luma.com/085nyxmj

General purpose agents like Claude Code and Manus use remarkably few tools. How? By giving agents access to a computer. With bash and filesystem tools, agents can perform actions without needing specialized bound tools for every task. Skills also offer two key advantages over traditional tools: 1. Token efficiency 2. Reduced cognitive load We’ve applied these principles and released skills for the Deep Agents CLI, available now. Here’s how you can use skills with Deep Agents: https://lnkd.in/dMUHA2hf

Agent skills are now available in the Deep Agents CLI, enabling you to use the large and growing collection of public skills with your agents. In this video we discuss: ✅ What agent skills are and why they’re interesting ✅ How agents make use of skills ✅ How you can use skills in the Deep Agents CLI https://lnkd.in/gJPDHRsW