--- title: Browser Use in Managed Deep Agents | daily.dev description: LangChain's Managed Deep Agents can now control a browser, letting agents click, scroll, and fill forms on sites without APIs. The browser runs in its own... canonical: https://daily.dev/posts/browser-use-in-managed-deep-agents-nmykxkgka twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: Browser Use in Managed Deep Agents | daily.dev og:description: LangChain's Managed Deep Agents can now control a browser, letting agents click, scroll, and fill forms on sites without APIs. The browser runs in its own... og:url: https://daily.dev/posts/browser-use-in-managed-deep-agents-nmykxkgka og:image: https://api.daily.dev/og/posts/nmYKXKGkA.png og:image:alt: Browser Use in Managed Deep Agents og:image:width: 1200 og:image:height: 630 og:locale: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # Browser Use in Managed Deep Agents **[LangChain](https://daily.dev/sources/langchain)** · 7 min read · 1 upvotes · 0 comments ## Summary LangChain's Managed Deep Agents can now control a browser, letting agents click, scroll, and fill forms on sites without APIs. The browser runs in its own LangSmith sandbox and supports thread-scoped or user-scoped sessions, with a handoff flow so a human can sign in without exposing passwords to the model. Core tools include browser_snapshot, browser_run, browser_screenshot, and browser_handoff, built on Browserbase's Stagehand SDK. On 100 WebVoyager tasks, the toolset passed 90% for $1.98 total, the highest pass rate among seven toolsets tested, beating Browserbase's MCP server (89% at 9x the cost). It also supports end-to-end testing by bridging the browser to a localhost app in the agent's sandbox. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### How does LangChain's Managed Deep Agents browser tool compare to Browserbase's MCP server on cost and accuracy? On 100 WebVoyager tasks, Managed Deep Agents' browser tools passed 90% of tasks for $1.98 total, the highest pass rate among seven toolsets tested. Browserbase's MCP server came closest with an 89% pass rate but cost 9 times as much. The tools are built on Stagehand, Browserbase's open source SDK for browser agents. _Comparing browser agent toolsets on cost and accuracy is easier with benchmarks like this on daily.dev._ ### How do I let an AI agent log into a website without giving it my password? Managed Deep Agents calls a browser_handoff tool when the agent hits a login page, giving a reason like needing the user to sign in. The user gets an interactive browser card in LangSmith Studio, signs in themselves, then clicks 'Hand back' so the agent continues already authenticated, without the password ever reaching the model. _Developers wiring up agent authentication flows track patterns like this via daily.dev._ ### What is the difference between thread-scoped and user-scoped browser sessions in LangChain Managed Deep Agents? Thread scope (the default) gives each conversation thread a fresh browser with no leftover cookies or logins. User scope shares one browser's cookies across a signed-in user's threads while still giving each thread its own window, so a login in one thread carries over to later threads instead of requiring a repeat sign-in. _Choosing the right session scope for stateful agent workflows is the kind of decision daily.dev helps developers track._ ## Similar posts on daily.dev - [GitHub - theredsix/agent-browser-protocol: Web browsing is continuous and async. Agents think in tools and steps. ABP reformats web navigation into the discrete, multimodal chat format agents know and](https://daily.dev/posts/github---theredsix-agent-browser-protocol-web-browsing-is-continuous-and-async-agents-think-in-too-kpdxygici) · Hacker News · 2 upvotes · 0 comments - [Browser Run: give your agents a browser](https://daily.dev/posts/browser-run-give-your-agents-a-browser-ivmgnoolf) · Cloudflare · 0 upvotes · 0 comments - [Self-Verifying AI Agents: Vercel's Agent-Browser in the Ralph Wiggum Loop](https://daily.dev/posts/self-verifying-ai-agents-vercel-s-agent-browser-in-the-ralph-wiggum-loop-3yew6qvxd) · Pulumi · 2 upvotes · 0 comments - [The best macOS browsers for AI agents, and how to use them](https://daily.dev/posts/the-best-macos-browsers-for-ai-agents-and-how-to-use-them-to4zrepri) · Flavio Copes · 2 upvotes · 1 comments --- Tags: [#ai-agents](https://daily.dev/tags/ai-agents), [#automation](https://daily.dev/tags/automation), [#langchain](https://daily.dev/tags/langchain) [View this post on daily.dev](https://daily.dev/posts/browser-use-in-managed-deep-agents-nmykxkgka) ```json {"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]} {"@context":"https://schema.org","@type":"TechArticle","headline":"Browser Use in Managed Deep Agents","url":"https://daily.dev/posts/browser-use-in-managed-deep-agents-nmykxkgka","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/browser-use-in-managed-deep-agents-nmykxkgka"},"datePublished":"2026-10-29T15:05:52.184Z","dateModified":"2026-10-29T17:15:45.533Z","description":"LangChain's Managed Deep Agents can now control a browser, letting agents click, scroll, and fill forms on sites without APIs. The browser runs in its own...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/8a4abc02932f5bb5d675e795754de7af?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/8a4abc02932f5bb5d675e795754de7af?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"LangChain","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"LangChain","logo":"https://media.daily.dev/image/upload/t_logo,f_auto/v1/logos/0f4f2629e0724a849823f8cd0d913e13","url":"https://daily.dev/sources/langchain"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/browser-use-in-managed-deep-agents-nmykxkgka","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":1},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"ai-agents,automation,langchain","timeRequired":"PT7M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"LangChain","item":"https://daily.dev/sources/langchain"},{"@type":"ListItem","position":3,"name":"Browser Use in Managed Deep Agents"}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/browser-use-in-managed-deep-agents-nmykxkgka#faq","mainEntity":[{"@type":"Question","name":"How does LangChain's Managed Deep Agents browser tool compare to Browserbase's MCP server on cost and accuracy?","acceptedAnswer":{"@type":"Answer","text":"On 100 WebVoyager tasks, Managed Deep Agents' browser tools passed 90% of tasks for $1.98 total, the highest pass rate among seven toolsets tested. Browserbase's MCP server came closest with an 89% pass rate but cost 9 times as much. The tools are built on Stagehand, Browserbase's open source SDK for browser agents. Comparing browser agent toolsets on cost and accuracy is easier with benchmarks like this on daily.dev."}},{"@type":"Question","name":"How do I let an AI agent log into a website without giving it my password?","acceptedAnswer":{"@type":"Answer","text":"Managed Deep Agents calls a browser_handoff tool when the agent hits a login page, giving a reason like needing the user to sign in. The user gets an interactive browser card in LangSmith Studio, signs in themselves, then clicks 'Hand back' so the agent continues already authenticated, without the password ever reaching the model. Developers wiring up agent authentication flows track patterns like this via daily.dev."}},{"@type":"Question","name":"What is the difference between thread-scoped and user-scoped browser sessions in LangChain Managed Deep Agents?","acceptedAnswer":{"@type":"Answer","text":"Thread scope (the default) gives each conversation thread a fresh browser with no leftover cookies or logins. User scope shares one browser's cookies across a signed-in user's threads while still giving each thread its own window, so a login in one thread carries over to later threads instead of requiring a repeat sign-in. Choosing the right session scope for stateful agent workflows is the kind of decision daily.dev helps developers track."}}]} ```