# agentsite-static-version=3940ceb1-dc74-4638-a88c-dce5a1c88d1d
# langchain.com

> AI-optimized mirror of langchain.com containing 99 pages totalling 119,969 words of clean markdown content, structured data, and semantic HTML. Original source: https://langchain.com. Last updated: 2026-09-17T02:49:43.855Z.

Each page is available as HTML (with JSON-LD structured data) and Markdown (text-only, ideal for LLMs and RAG).

## Homepage

- [LangChain: the open agent platform to own your intelligence](/content/site-root.html): LangChain enables every company to own their intelligence. Control, govern, and compound intelligence with an open agent engineering platform. (479 words)

## Articles & Blog Posts

- [LangChain Events: AI Conferences, Meetups & Webinars](/content/events/index.html): Find upcoming LangChain events, community meetups, and live webinars. Connect with AI engineers and stay current on the latest in LLM and agent development. (1,153 words)
- [Terms of Service](/content/terms-of-service/index.html) (5,612 words)
- [The 6 Best AI Agent Sandboxes for Isolated Code Execution](/content/resources/best-agent-sandboxes/index.html): Compare six AI agent sandboxes by isolation, credentials, egress, state, deployment model, and lifecycle cost. (4,336 words)
- [Customer Experience (CX) Agents in Production: Lessons from Lyft, Vodafone, and LATAM Airlines](/content/resources/customer-experience-cx-agents-in-production/index.html): Customer Experience (CX) Agents in Production: Lessons from Lyft, Vodafone, and LATAM Airlines (3,814 words)
- [Building Governed Agents: A Framework for Cost, Control, and Compliance.](/content/resources/building-governed-agents-a-framework-for-cost-control-and-compliance.html): Agents are becoming part of production infrastructure. They answer customer questions, write and deploy code, retrieve company knowledge, and take action across business systems. As their autonomy grows, the central governance question is how policy can be enforced on every interaction across models, data, tools, and providers without slowing adoption. (2,901 words)
- [Evaluating AI Agents at the Run, Trace, and Thread Level](/content/resources/agent-evals/index.html): Learn how to evaluate AI agents end-to-end, including trajectory-based evals, offline vs. online testing, LLM judges, and human review to spot drift and failures. (2,878 words)
- [AI Agent Monitoring: Why Traditional APM Falls Short](/content/resources/ai-agent-monitoring/index.html): Traditional APM can’t catch every agent failure, and how tracing, cost attribution, and evaluation loops help you debug and improve agent performance. (2,986 words)
- [AI Observability in the Agent Development Lifecycle](/content/resources/ai-observability/index.html): Understand AI observability beyond uptime: how traces explain agent behavior, where failures show up, and how monitoring supports evals and fixes. (3,118 words)
- [Support Plans](/content/support-plans/index.html) (1,661 words)
- [LangChain State of AI Agents Report: 2024 Trends](/content/stateofaiagents/index.html): Are AI agents being used in production? What's the biggest challenge to deploying agents - cost, quality, skill, or latency? Get insights on AI agent adoption and sentiment for devs and enterprises today. (1,830 words)
- [Open Source Agent Stack | LangChain OSS Overview](/content/oss-overview/index.html): Build agents with proven patterns using Deep Agents, LangChain, and LangGraph. Start fast with high-level abstractions or drop down for full control. (673 words)
- [The best AI agent frameworks in 2026](/content/resources/ai-agent-frameworks/index.html): We reviewed 7 AI agent frameworks across orchestration, observability, and production readiness. See how LangGraph, CrewAI, Microsoft Agent Framework, and others compare. (2,133 words)
- [LangChain Community: Connect, Learn & Build AI Agents](/content/community/index.html): At LangChain, community is at the heart of what we do. Join us in sharing insights and driving the future of AI development together. (431 words)
- [LangGraph AI Agent Framework for Production Applications](/content/built-with-langgraph/index.html): Read the customer stories from companies that choose LangGraph to build their production GenAI applications. (258 words)
- [LLM Gateway for AI Agent & Model Governance | LangSmith](/content/langsmith/llm-gateway/index.html): Centralize model routing, control spend, and ensure reliability. LangSmith LLM Gateway provides model fallbacks, spending caps, and real-time cost monitoring for your agents. (637 words)
- [LangSmith: Agent & LLM Observability Platform](/content/langsmith/observability/index.html): Complete AI agent and LLM observability platform with tracing and real-time monitoring. Debug agents, find failures fast, and track costs and latency. (808 words)
- [Brand Assets - LangChain Logos & Usage Guidelines](/content/brand-assets/index.html): Download official LangChain, LangSmith, and open source framework logos. Light and dark variants for documentation, presentations, and community content. (325 words)
- [LangSmith: AI Agent & LLM Observability and Evals Platform](/content/langsmith-platform/index.html): LangSmith is the complete framework agnostic AI agent and LLM observability, evaluation, and deployment platform. (495 words)
- [LangChain Agent Inbox Case Study: Superhuman's AI](/content/breakoutagents/superhuman/index.html): See how Superhuman built an AI-powered search assistant for users to quickly navigate their inbox and calendar. Learn about their UX, evoliution of agent architecture, and prompt engineering., Stream (1,130 words)
- [Built with LangGraph](/content/resources/built-with-langgraph/index.html) (39 words)
- [Community v2](/content/v-2/community-v2-sdfasdfasdf-asdfasdfafsd/index.html): At LangChain, community is at the heart of what we do. Join us in sharing insights and driving the future of AI development together. (362 words)
- [dcode: Open-Source Terminal Coding Agent | Deep Agents Code](/content/dcode/index.html): dcode is an open-source terminal coding agent built on Deep Agents SDK. Customize your agent harness, control code execution, and bring your own model. (550 words)
- [LangSmith: Agent Deployment Infrastructure for Production AI Agents](/content/langsmith/deployment/index.html): LangSmith Deployment provides purpose-built infrastructure for deploying and managing long-running agent workloads. (351 words)
- [AI Agent Case Study: Ramp's Tour Guide for Financial Ops](/content/breakoutagents/ramp/index.html): See how Ramp built an AI tour guide that helps users navigate Ramp's platform for financial operations. Learn about their UX, agent architecture, prompt engineering, and more., Stream (1,161 words)
- [LangChain Partner Network](/content/langchain-partner-network/index.html): Solve your biggest AI challenges with LangChain’s network of cloud and services partners. (143 words)
- [Replit Agent Case Study: AI Agent Architecture & Build](/content/breakoutagents/replit/index.html): See how Replit built an agent that can build software apps from scratch to deployment. Learn about their agent architecture, UX choices, and prompt engineering techniques. (1,175 words)
- [Deep Agents: Open Source Agent Harness | LangChain](/content/deep-agents/index.html): Deep Agents handles planning, context management, and subagent orchestration, so agents can run long, complex work like research and coding. (435 words)
- [LangSmith for Startups](/content/startups/index.html): Up to $10K in LangSmith credits, discounted pricing, and founder programming for early-stage startups. (481 words)
- [LangSmith Engine](/content/langsmith/engine/index.html): LangSmith Engine analyzes production traces, groups related failures, and recommends fixes so your team can improve agent quality faster. (466 words)
- [AI Sandboxes: Secure Runtime for Agent Code | LangSmith](/content/langsmith/sandboxes/index.html): Ephemeral, isolated sandboxes for agent-generated code. MicroVM isolation, sub-second starts, stateful sessions. Run untrusted code safely at scale. (587 words)
- [SmithDB Early Access Waitlist](/content/smithdb-early-access-waitlist/index.html): Sign up to request access to SmithDB early access waitlist. SmithDB is LangChain's purpose-built distributed database for agent observability. (146 words)
- [Langsmith LLM Gateway Data Protection Request](/content/langsmith-llm-gateway-data-protection-request/index.html): Opt into Data Protection policies for LangSmith LLM Gateway (137 words)
- [AI Answer Engine Case Study: Perplexity Pro Search](/content/breakoutagents/perplexity/index.html): See how Perplexity built an AI answer engine that lets users answer complex query searches like a Pro. Learn about their UX, agent design, and prompt engineering techniques. (1,019 words)
- [LangChain Slack Community Code of Conduct & Values](/content/community-code/index.html): This Code of Conduct is applicable to all LangChain Community spaces — whether online or offline. (986 words)
- [LangSmith Tuned Evaluators: Custom Evaluator Request](/content/langsmith-tuned-evaluators-custom-evaluator-request.html): Request a custom LangSmith Tuned Evaluator. They help teams find agent conversations that need attention, understand what went wrong, and turn those traces into useful examples for improving their agents. (140 words)
- [LangSmith Plans and Pricing](/content/pricing/index.html): Pricing for LangChain products for teams of any size. Choose the plan that suits your needs, whether you're an individual developer or enterprise. (514 words)
- [How we built Agent Builder’s memory system](/content/resources/agent-builder-memory/index.html) (66 words)
- [LangSmith LLM Gateway Self-hosted Access Request](/content/langsmith-llm-gateway-self-hosted-access-request/index.html): Self-hosted access request for LangSmith LLM Gateway. LLM Gateway is a central governance layer that sits between your agents and the models they call. It gives teams one place to enforce runtime controls across agents, models, and providers, helping them consistently govern model usage while avoiding vendor lock-in. (124 words)
- [Designing Agents](/content/resources/designing-agents/index.html) (33 words)
- [LangChain Community Slack for GenAI Developers](/content/join-community/index.html): Sign up for the LangChain Community Slack to chat with other developers and ask questions about building GenAI applications. (802 words)
- [Breakout Agentic Apps](/content/resources/breakout-agentic-apps/index.html) (39 words)
- [LangChain Partner Network Form](/content/langchain-partner-network-form/index.html): Solve your biggest AI challenges with LangChain’s network of cloud and services partners. (387 words)
- [Templates](/content/templates/index.html): Templates (54 words)
- [Agent observability powers agent evaluation](/content/resources/agent-observability-powers-agent-evaluation/index.html) (36 words)
- [LangChain Resources: Guides, Reports & AI Agent Insights](/content/resources/index.html): Curated content for the AI engineer developing their agent or LLM application. Get the latest on AI trends and learn best practices. (90 words)
- [Agentic AI Apps: Breakout Case Studies | LangChain](/content/breakoutagents/index.html): Dive into the stories of companies pushing the boundaries of AI agents. Learn "why" and "how" they made specific architecture, UX, prompt engineering, and evaluation choices for high-impact results. (106 words)
- [Contact the LangChain Sales Team](/content/contact-sales/index.html): You can expect a conversation with a LangSmith expert to assess if LangSmith fits your company's needs. (63 words)
- [How and when to build multi-agent systems](/content/blog/how-and-when-to-build-multi-agent-systems/index.html): Learn when to build multi-agent AI systems and when to avoid them. Context engineering best practices, reliability patterns, and production insights. (1,537 words)
- [Privacy policy](/content/privacy-policy/index.html) (4,018 words)
- [How Candidly Built State-Aware Agent Harnesses in LangSmith](/content/blog/how-candidly-built-state-aware-agent-harnesses-with-langsmith.html): Candidly's agent Cait reads partial traces to infer user state mid-conversation and steer replies, using a LangSmith labeling pipeline at 92.3% human agreement. (1,177 words)
- [9 LLM Observability Tools for Production AI Agents](/content/resources/llm-observability-tools/index.html): Compare the best LLM observability tools for production AI agents, with practical tradeoffs for traces, evals, debugging, cost, and review workflows. (5,276 words)
- [♠️ SPADE: Automatically Digging up Evals based on Prompt Refinements](/content/blog/spade-automatically-digging-up-evals-based-on-prompt-refinements.html): Generate evaluation functions for your LLM prompts with SPADE. Analyze prompt refinements to discover useful evals and improve chain reliability. (1,378 words)
- [How Athena Intelligence optimized research reports with LangSmith, LangChain, and LangGraph](/content/blog/customers-athena-intelligence/index.html): See how an AI-powered employee for enterprise analytics used the LangSmith playground and debugging features to quickly identify LLM issues and to generate complex research reports. (918 words)
- [How to Build a Custom Agent Harness](/content/blog/how-to-build-a-custom-agent-harness/index.html): Effective agents are built with harnesses that are tightly coupled with the task at hand. The easiest way to build a custom harness is with LangChain's create_agent plus middleware. This guide covers the core agent loop and how you can customize it for your agent's use case. (1,107 words)
- [LangChain and LangGraph Agent Frameworks Reach v1.0 Milestones](/content/blog/langchain-langgraph-1dot0/index.html): LangChain 1.0 and LangGraph 1.0 are here. Build production-ready AI agents faster with standardized tools, middleware customization, and durable state. (1,677 words)
- [Why We Rebuilt LangChain’s Chatbot and What We Learned](/content/blog/rebuilding-chat-langchain/index.html): Learn how LangChain rebuilt their chatbot using Deep Agents and subgraphs for sub-15-second responses with precise citations—and what you can apply. (3,194 words)
- [Few-shot prompting to improve tool-calling performance](/content/blog/few-shot-prompting-to-improve-tool-calling-performance.html): We ran a few experiments, which show how few-shot prompting can significantly enhance model accuracy - especially for complex tasks. Read on for how we did it (and the results). (1,987 words)
- [LangGraph: Multi-Agent Workflows](/content/blog/langgraph-multi-agent-workflows/index.html): Build multi-agent AI workflows with LangGraph. Create specialized agents with unique prompts and tools, then connect them for better LLM results. (1,271 words)
- [Improving Deep Agents with harness engineering](/content/blog/improving-deep-agents-with-harness-engineering/index.html): Harness engineering improved LangChain's coding agent from Top 30 to Top 5 on Terminal Bench using self-verification, tracing, and context optimization. (1,748 words)
- [Reflection Agents](/content/blog/reflection-agents/index.html): Reflection is a prompting strategy used to improve the quality and success rate of agents and similar AI systems. This post outlines how to build 3 reflection techniques using LangGraph, including implementations of Reflexion and Language Agent Tree Search. (1,296 words)
- [AI Agent Observability: Tracing, Testing, and Improving Agents](/content/resources/agent-observability/index.html): Go beyond logs and print statements. Learn how to trace, test, and improve AI agents in production with step-by-step observability and automated evaluation. (2,684 words)
- [Multi-Vector Retriever for RAG on tables, text, and images](/content/blog/semi-structured-multi-modal-rag/index.html): Learn how to implement multi-vector retriever for RAG across tables, text, and images. Explore cookbooks for semi-structured and multi-modal data retrieval. (1,096 words)
- [LangMem SDK for agent long-term memory](/content/blog/langmem-sdk-launch/index.html): Build smarter AI agents with LangMem SDK's long-term memory. Extract insights, optimize behavior, and personalize experiences over time. (1,452 words)
- [Deep Agents v0.5](/content/blog/deep-agents-v0-5/index.html): 💡TL;DR: We’ve released new minor versions of deepagents & deepagentsjs, featuring async (non-blocking) subagents, expanded multi-modal filesystem support, and more. See the changelog for details. Async subagents. Deep Agents can now delegate work to remote agents that run in the background. As opposed to the existing inline subagents, which (947 words)
- [How we built LangChain’s GTM Agent](/content/blog/how-we-built-langchains-gtm-agent/index.html): Learn how we built a GTM agent that increased lead conversion by 250% while saving each sales rep 40 hours per month (1,358 words)
- [LangChain Careers: Enabling companies to own their intelligence](/content/careers/index.html): We're a growing team of builders helping every company own their intelligence, making an outsized impact along the way., LangChain Jobs (1,935 words)
- [Sandboxes](/content/resources/agents-need-their-own-computer/index.html): Agents need their own computer. Here's how to give them one safely. (2,396 words)
- [The two patterns by which agents connect sandboxes](/content/blog/the-two-patterns-by-which-agents-connect-sandboxes/index.html): Learn two architecture patterns for integrating AI agents with sandboxes: running agents inside vs. using sandboxes as tools. Compare trade-offs. (1,366 words)
- [The Agent Development Lifecycle: Build, Test, Deploy & Monitor AI Agents | LangChain](/content/blog/the-agent-development-lifecycle/index.html): Learn how leading engineering teams ship AI agents reliably and repeatedly using a four-phase agent development lifecycle: Build, Test, Deploy, and Monitor. Includes guidance on evals, runtimes, observability, and governance at scale. (2,719 words)
- [Debugging Deep Agents with LangSmith](/content/blog/debugging-deep-agents-with-langsmith/index.html): Debug deep agents with LangSmith's tracing and analysis. Analyze complex traces, optimize prompts with Polly, and improve performance. (1,203 words)
- [Open Source Extraction Service](/content/blog/open-source-extraction-service/index.html) (1,414 words)
- [How Auth Proxy secures LangSmith agent sandboxes](/content/blog/how-auth-proxy-secures-network-access-for-langsmith-agent-sandboxes.html): Agents need credentials and network access to do useful work, but those capabilities create new security risks. This post explains how Auth Proxy keeps secrets out of LangSmith Sandboxes runtimes, constrains agent egress, and gives teams infrastructure-level control over how agents reach external services. (1,696 words)
- [Agent observability needs feedback to power learning](/content/blog/agent-observability-needs-feedback-to-power-learning.html): Traces show what an agent did; feedback shows what it meant. How explicit, implicit, LLM-as-judge and rule-based feedback turn observability into learning. (1,454 words)
- [How LangSmith and LangChain OSS Help You Meet EU AI Act Requirements](/content/blog/langsmith-langchain-oss-eu-ai-act/index.html): The EU AI Act compliance deadline is August 2, 2026. Learn what the EU AI Act requires, and how LangSmith and LangChain OSS products help you meet each requirement. (1,213 words)
- [Agentic Engineering: How Swarms of AI Agents Are Redefining Software Engineering](/content/blog/agentic-engineering-redefining-software-engineering.html): Multi-agent systems that mirror real engineering teams — not just code faster — can cut debug time by 93% and compress cross-team delivery. Here's the architecture built on LangGraph. (2,269 words)
- [State of AI Agents](/content/state-of-agent-engineering/index.html): LangChain provides the engineering platform and open source frameworks developers use to build, test, and deploy reliable AI agents. (1,592 words)
- [Interpreter Skills: Building Workflows for Agents](/content/blog/interpreter-skills/index.html): Interpreter skills extend agent skills with a TypeScript module the agent can import and run. This lets you build more capable workflows with your agents. (1,064 words)
- [How Minimal built a multi-agent customer support system with LangGraph & LangSmith](/content/blog/how-minimal-built-a-multi-agent-customer-support-system-with-langgraph-langsmith.html): Learn how Minimal built a multi-agent AI system with LangGraph & LangSmith to automate 90% of e-commerce support tickets, delivering 80%+ efficiency gains. (824 words)
- [Fleet](/content/langsmith/fleet/index.html): LangSmith Fleet enables anyone to build powerful agents using natural language. (575 words)
- [Syncing data sources to vector stores](/content/blog/syncing-data-sources-to-vector-stores/index.html): Sync vector stores with LangChain's Indexing API. Avoid duplicate content, reduce costs, and efficiently manage document updates with automated cleanup. (862 words)
- [Managed Deep Agents: the fastest way to ship a production deep agent](/content/blog/introducing-managed-deep-agents/index.html): Run deep agents in production with durable execution, sandboxes, tool access, and LangSmith observability, without building the runtime yourself. Now in private beta (1,135 words)
- [Benchmarking Multi-Agent Architectures](/content/blog/benchmarking-multi-agent-architectures/index.html): Benchmarks comparing single agent, swarm, and supervisor multi-agent architectures. LangGraph's supervisor improvements achieved 50% performance gains. (1,716 words)
- [Evaluating Deep Agents CLI on Terminal Bench 2.0](/content/blog/evaluating-deepagents-cli-on-terminal-bench-2-0/index.html): Evaluate Deep Agents CLI on Terminal Bench 2.0. This open-source coding agent achieves 42.5% accuracy using isolated sandbox environments. (781 words)
- [LangChain Customer Stories](/content/customers/index.html): Read the customer stories from companies that choose LangChain, LangSmith, and LangGraph to build their needle-moving GenAI applications. (594 words)
- [OpenWiki 0.2 brings OKF to codebase documentation](/content/blog/openwiki-0-2-adds-okf-support/index.html): OpenWiki 0.2 generates codebase wikis in the OKF format, helping developers organize repo docs with metadata, changelogs, and agent-friendly retrieval. (749 words)
- [Introducing Pytest and Vitest integrations for LangSmith Evaluations](/content/blog/pytest-and-vitest-for-langsmith-evals/index.html): Introducing a new way to run evals using LangSmith’s Pytest and Vitest/Jest integrations. (1,192 words)
- [LangSmith: AI Agent & LLM Model Evaluation Platform](/content/langsmith/evaluation/index.html): Test prompts, monitor quality, debug agent failures with our LLM & AI agent evaluation platform. Catch regressions before users do. (686 words)
- [LangChain Blog](/content/blog/index.html): Explore tutorials, case studies, and technical insights on building AI agents with LangSmith, Deep Agents, LangGraph, and LangChain. Learn from experts. (282 words)
- [How to Choose the Right Sandbox for AI Agents](/content/blog/how-to-choose-the-right-sandbox-for-your-agent/index.html): Learn how to choose a secure sandbox for AI agents, with guidance on filesystem isolation, network access, resource limits, and microVMs. (1,177 words)
- [In software, the code documents the app. In AI, the traces do.](/content/blog/in-software-the-code-documents-the-app-in-ai-the-traces-do.html): Discover how traces for documentation replace code as the source of truth for AI agents. Learn to debug, test, and optimize agent behavior effectively. (1,426 words)
- [How Klarna's AI assistant redefined customer support at scale for 85 million active users](/content/blog/customers-klarna/index.html): Klarna's AI assistant is revolutionizing the personal shopping experience, including customer service and productivity. See how they used LangGraph and LangSmith to achieve 80% faster customer resolution times. (474 words)
- [Give your agent its own computer](/content/blog/give-your-ai-agent-its-own-computer/index.html): Running code execution in an AI agent is harder than it looks. Your agent needs a real computer (filesystem, shell, package manager, persistent state) but handing it access to your infrastructure is dangerous.Think about it this way: you use one laptop. You are n of one. But agents are going to run millions of tasks, and each one needs its own computer to work from. That's the infrastructure shift happening right now. Satya Nadella put it plainly: "Every agent needs a computer." The question is what that computer looks like, and how you give it to them safely.LangSmith Sandboxes are our answer to that. Here's why it matters, and why doing it yourself is harder than it sounds. (1,371 words)
- [LangGraph: Agent Orchestration Framework for Reliable AI Agents](/content/langgraph/index.html): Control agent workflows with LangGraph: durable execution, memory, streaming, and human-in-the-loop. (406 words)
- [OpenWiki: Open Source Repo Documentation for Coding Agents](/content/blog/introducing-openwiki-an-open-source-agent-for-repo-documentation.html): OpenWiki generates and maintains codebase documentation so coding agents can find the repo context they need without loading everything into one instruction file. (734 words)
- [Wiki Memory: File-Based Memory for AI Agents | LangChain](/content/blog/wiki-memory/index.html): Wiki memory uses an agent to compress raw data into a persistent, file-based knowledge base. How it differs from RAG, real examples, and when to use it. (656 words)
- [LangChain: Open Source AI Agent Framework for Any Model](/content/langchain/index.html): Build agents in minutes with LangChain, the open source framework with pre-built agent architectures and 1,000+ integrations. (277 words)
- [New in Fleet: Deploy AI agents to Slack in one click](/content/blog/new-in-langsmith-fleet-bring-agents-into-slack-in-one-click.html): Build custom AI agents in Fleet without code, then deploy them to Slack in one click. Give agents custom identities, use them in channels and threads, and keep work moving where your team already collaborates. (775 words)

## About Pages

- [About LangChain: an open platform to own your intelligence](/content/about/index.html): LangChain is the platform companies use to own their agent intelligence, from open source frameworks to production-ready infrastructure. (685 words)

## Resources

- [Full Page Index](/index.html): Browse all cached pages with rich metadata
- [About This Cache](/about.html): Methodology, technical details, and usage guidelines
- [XML Sitemap](/sitemap.xml): Machine-readable sitemap for crawler discovery
- [Robots.txt](/robots.txt): Crawler directives
- [AgentSite Network](https://agentsite.network/network.html): Public index of AgentSites and their machine-readable resources