{
 "sym": "Bu",
 "updated": "2026-08-06",
 "verdict": "Claude in Chrome for the human-in-the-loop case \u2014 it is the only assisted-browsing agent that ships published adversarial safety numbers (prompt-injection success cut 23.6%\u219211.2%, and 35.7%\u21920% on four browser-specific attacks) and org-level site allowlists, on every paid Claude plan. Browser-use when you are building the automation yourself \u2014 107k GitHub stars, MIT, and the de-facto open-source standard. Browserbase + Stagehand when that automation has to run in production at volume (proxies, captchas, session replay). Comet when you want a free consumer agentic browser; Playwright MCP when the browser work is your own app \u2014 testing, verification, coding-agent loops \u2014 where it is free and deterministic. The category consolidated hard in mid-2026: OpenAI is retiring Atlas (Aug 9), Google folded Mariner into Gemini-in-Chrome, Microsoft folded Copilot Mode into Edge. The standalone AI browser is dying; the browser agent is not.",
 "top5": [
  {
   "rank": 1,
   "name": "Claude in Chrome",
   "maker": "Anthropic",
   "url": "https://claude.com/claude-for-chrome",
   "docs": "https://support.claude.com/en/articles/12012173-getting-started-with-claude-for-chrome",
   "pricing": "Included in paid Claude plans (Pro $17\u201320/mo \u00b7 Max $100\u2013200/mo \u00b7 Team/Enterprise) \u00b7 beta",
   "best_for": "Founders and teams who want an agent working inside their real, logged-in browser \u2014 inbox triage, form filling, dashboard extraction, scheduled recurring tasks \u2014 with admin-managed site controls.",
   "why": "The safety-engineering leader in a category defined by prompt-injection risk: Anthropic published red-team numbers (123 cases, 29 attack scenarios; success rate 23.6%\u219211.2% after mitigations, 35.7%\u21920% on four browser-specific attacks) and shipped org-wide allowlists/blocklists when it expanded from Max-only to all paid plans (Dec 18, 2025). It is also the best-integrated: hands for Claude Code (build in terminal, verify in browser) and Cowork, plus multi-tab work, background tasks, and cron-style scheduled browsing.",
   "watch": "Still labeled beta with 'unique risks' \u2014 Anthropic itself tells you to keep it away from financial transactions and passwords. Chrome-desktop-only, no API surface: this is an assistant, not infrastructure you can build a product on.",
   "evidence": [
    {
     "stat": "Prompt-injection attack success 23.6%\u219211.2% after mitigations; 35.7%\u21920% on four browser-specific attack types (Aug 2025 red-team, 123 cases)",
     "src": "https://claude.com/blog/claude-for-chrome"
    },
    {
     "stat": "Expanded from 1,000-user Max pilot (Aug 2025) to all Max (Nov 24, 2025) to Pro/Team/Enterprise with admin site allowlists (Dec 18, 2025)",
     "src": "https://claude.com/blog/claude-for-chrome"
    },
    {
     "stat": "Scheduled recurring browser tasks, multi-tab groups, background runs, Claude Code integration (verified Aug 2026)",
     "src": "https://support.claude.com/en/articles/12012173-getting-started-with-claude-for-chrome"
    }
   ],
   "tile_note": "assisted browsing"
  },
  {
   "rank": 2,
   "name": "Browser-use",
   "maker": "Browser Use (YC W25, MIT)",
   "url": "https://browser-use.com",
   "docs": "https://docs.browser-use.com",
   "pricing": "OSS free (MIT) + your LLM tokens \u00b7 Cloud: browsers $0.02/hr metered by minute, hosted agents from $0.36/M input \u00b7 $1.44/M output tokens",
   "best_for": "Engineers building their own browser automation \u2014 the open-source library that turns any LLM into a web agent, with a usage-priced cloud when you outgrow localhost.",
   "why": "The open-source standard: 107k GitHub stars (Aug 2026), MIT-licensed, Python + TypeScript, any model provider \u2014 it is the layer Manus and 20+ YC startups were built on. Its DOM-to-text approach (parse interactive elements instead of screenshots) made agents cheaper and more reliable than vision-first designs, and the repo claims #1 on the Odysseys long-horizon benchmark (87.4% across 200 web tasks). $17M seed led by Felicis (Mar 2025) with Paul Graham participating.",
   "watch": "Quality is whatever model you plug in, and long agent runs burn tokens fast. The cloud business is young and the space is crowded with better-capitalized infra (Browserbase); benchmark claims are self-reported.",
   "evidence": [
    {
     "stat": "107.1k GitHub stars, MIT license, ~9.9k commits (Aug 2026)",
     "src": "https://github.com/browser-use/browser-use"
    },
    {
     "stat": "$17M seed led by Felicis, with Paul Graham, A Capital, Nexus (Mar 23, 2025)",
     "src": "https://techcrunch.com/2025/03/23/browser-use-the-tool-making-it-easier-for-ai-agents-to-navigate-websites-raises-17m/"
    },
    {
     "stat": "Self-reported #1 on Odysseys leaderboard: 87.4% avg across 200 long-horizon web tasks (repo, Aug 2026)",
     "src": "https://github.com/browser-use/browser-use"
    }
   ],
   "tile_note": "open-source control"
  },
  {
   "rank": 3,
   "name": "Browserbase + Stagehand",
   "maker": "Browserbase",
   "url": "https://www.browserbase.com",
   "docs": "https://docs.browserbase.com",
   "pricing": "Free tier \u00b7 Developer $20/mo (100 browser-hrs, $0.12/hr after) \u00b7 Startup $99/mo \u00b7 Enterprise custom; Stagehand SDK free (MIT)",
   "best_for": "Production browser automation at scale \u2014 headless browser fleets with proxies, captcha solving, session replay, and Stagehand's self-healing act/extract/observe primitives on top.",
   "why": "The infrastructure pick: $40M Series B at a $300M valuation (Jun 2025), ~50M sessions served in 2025, and marquee production users \u2014 Airtable built its Hyperagent platform on it (Jul 2026), Amplitude runs sales automation on it. Stagehand (MIT, ~23k stars, TypeScript + Python) is the most credible middle path between brittle Playwright selectors and fully autonomous agents, and the Jun 30, 2026 'Browserbase Agents' launch added fully managed prompt-to-agent runs.",
   "watch": "Costs compound at volume (browser-hours + proxies + agent runs are metered separately). The captcha/anti-bot arms race means reliability on hostile sites is never guaranteed, and its new managed Agents put it in competition with its own customers.",
   "evidence": [
    {
     "stat": "$40M Series B at $300M valuation (Jun 2025); ~50M sessions in 2025",
     "src": "https://www.firecrawl.dev/blog/best-browser-agents"
    },
    {
     "stat": "Airtable's Hyperagent built on Browserbase (customer story, Jul 2, 2026); Browserbase Agents launched Jun 30, 2026",
     "src": "https://www.browserbase.com/blog"
    },
    {
     "stat": "Stagehand: MIT, act/extract/observe/agent primitives, TS + Python, self-healing automations (verified Aug 2026)",
     "src": "https://www.stagehand.dev/"
    }
   ],
   "tile_note": "production browser infra"
  },
  {
   "rank": 4,
   "name": "Perplexity Comet",
   "maker": "Perplexity",
   "url": "https://www.perplexity.ai/comet",
   "docs": "https://www.perplexity.ai/help-center/en/collections/18799298-comet.html",
   "pricing": "Free \u00b7 Comet Plus $5/mo \u00b7 Pro $20/mo \u00b7 Max $200/mo (background assistants) \u00b7 Enterprise from $40/seat",
   "best_for": "A free, full agentic browser for research-heavy individual work \u2014 summarize, compare across tabs, handle email, shop \u2014 on Mac, Windows, iOS, and Android.",
   "why": "The distribution winner among standalone AI browsers \u2014 and after Atlas's retirement, effectively the last one standing at scale: free worldwide since Oct 2025, Android (Nov 2025) and iOS (Mar 2026) apps, enterprise MDM deployment since Mar 2026. Its assistant works inside your logged-in sessions, which is exactly why Amazon sued \u2014 making Comet the test case for whether user agents may act in authenticated accounts.",
   "watch": "The Amazon CFAA litigation (preliminary injunction Mar 2026, stayed pending Ninth Circuit appeal) could constrain logged-in commerce agents for everyone. Security researchers keep landing hits (CometJacking; a sub-4-minute phishing demo, Mar 2026). Consumer product, no developer API.",
   "evidence": [
    {
     "stat": "Launched Jul 9, 2025 (Mac/Windows); free worldwide Oct 2025; Android Nov 20, 2025; iOS Mar 18, 2026",
     "src": "https://en.wikipedia.org/wiki/Comet_(browser)"
    },
    {
     "stat": "Amazon legal demand over agentic shopping in logged-in accounts (Nov 2025); Perplexity refused \u2014 'users have a right to employ AI agents'",
     "src": "https://www.perplexity.ai/hub/blog/bullying-is-not-innovation"
    },
    {
     "stat": "Free tier + Comet Plus $5/mo + Pro $20/mo + Max $200/mo with autonomous background assistants (2026 pricing)",
     "src": "https://www.eesel.ai/blog/perplexity-comet-pricing"
    }
   ],
   "tile_note": "the free consumer agent"
  },
  {
   "rank": 5,
   "name": "Playwright MCP",
   "maker": "Microsoft",
   "url": "https://github.com/microsoft/playwright-mcp",
   "docs": "https://github.com/microsoft/playwright-mcp#readme",
   "pricing": "Free, Apache-2.0 \u00b7 your LLM tokens only",
   "best_for": "Giving any MCP-speaking agent (Claude Code, Cursor, VS Code, Cline) deterministic browser hands \u2014 testing your own app, verifying agent-built UIs, automating cooperative sites.",
   "why": "The dev-loop default by sheer volume: 6.3M npm downloads per week (Aug 2026) and 35.5k stars. Its accessibility-snapshot approach \u2014 structured a11y trees instead of screenshots \u2014 is fast, cheap, and works without vision models, with 60+ tools covering navigation, network mocking, and trace recording. If your coding agent needs to see and click what it just built, this is the free, boring, correct answer.",
   "watch": "A tool server, not an agent \u2014 you bring the loop and the judgment. No stealth, proxies, or captcha handling, so it struggles on hostile third-party sites; snapshot-heavy sessions get token-expensive on long tasks (Microsoft itself notes CLI workflows are more token-efficient at high throughput).",
   "evidence": [
    {
     "stat": "6.30M weekly npm downloads (week of Jul 30\u2013Aug 5, 2026)",
     "src": "https://api.npmjs.org/downloads/point/last-week/@playwright/mcp"
    },
    {
     "stat": "35.5k GitHub stars, Apache-2.0, 60+ tools (Aug 2026)",
     "src": "https://github.com/microsoft/playwright-mcp"
    },
    {
     "stat": "Accessibility snapshots over screenshots \u2014 no vision model required, deterministic tool application",
     "src": "https://github.com/microsoft/playwright-mcp"
    }
   ],
   "tile_note": "the free dev default"
  }
 ],
 "matrix": {
  "cols": [
   "Form factor",
   "Pricing",
   "Open source",
   "Autonomy",
   "Logged-in sites",
   "Stealth / captcha",
   "Enterprise controls",
   "Best fit"
  ],
  "rows": [
   [
    "Claude in Chrome",
    "Extension in your Chrome",
    "Paid Claude plans",
    "No",
    "Supervised \u2192 autonomous",
    "Yes (your session)",
    "No",
    "Org allowlists/blocklists",
    "Personal + team ops"
   ],
   [
    "Browser-use",
    "Library + cloud API",
    "Free + tokens \u00b7 usage cloud",
    "Yes (MIT)",
    "Fully programmatic",
    "Yes (managed sessions)",
    "Cloud: proxies + captcha",
    "Minimal",
    "Build your own automation"
   ],
   [
    "Browserbase + Stagehand",
    "Cloud infra + SDK",
    "$20\u201399/mo + usage",
    "SDK yes (MIT)",
    "Programmatic + managed agents",
    "Yes (persisted contexts)",
    "Proxies \u00b7 captcha \u00b7 replay",
    "SOC2 \u00b7 HIPAA \u00b7 SSO",
    "Production at scale"
   ],
   [
    "Perplexity Comet",
    "Standalone browser",
    "Free \u00b7 Plus $5 \u00b7 Max $200",
    "No",
    "Assistant \u2192 background agents",
    "Yes (your session)",
    "No",
    "MDM (Mar 2026)",
    "Free consumer agent"
   ],
   [
    "Playwright MCP",
    "MCP server",
    "Free + tokens",
    "Yes (Apache)",
    "Tool server (agent brings loop)",
    "Yes (local profile)",
    "No",
    "No",
    "Testing + dev loops"
   ],
   [
    "Gemini in Chrome",
    "Built into Chrome",
    "AI Pro/Ultra sub",
    "No",
    "Auto-browse (preview)",
    "Yes (your session)",
    "No",
    "Workspace policy",
    "Google-stack consumers"
   ],
   [
    "Skyvern",
    "Cloud + OSS platform",
    "5k free credits \u00b7 usage",
    "Yes (AGPL)",
    "Vision-driven workflows",
    "Yes (2FA-capable)",
    "Yes",
    "SOC2 \u00b7 HIPAA",
    "Enterprise form workflows"
   ],
   [
    "ChatGPT agent (Atlas \u2020Aug 9)",
    "Browser \u2192 folding into ChatGPT",
    "Plus/Pro/Business",
    "No",
    "Agent mode (preview)",
    "Yes",
    "No",
    "Business tier",
    "ChatGPT loyalists"
   ]
  ]
 },
 "rules": [
  {
   "if": "The site has an API or MCP server",
   "then": "Use it and skip this element \u2014 browser agents are for the long tail without APIs. They are slower, costlier, and more brittle than any real integration, and the vendor may sue (see Amazon v. Perplexity)."
  },
  {
   "if": "You want an agent doing your browsing \u2014 inbox, dashboards, forms, recurring chores \u2014 under your eyes",
   "then": "Claude in Chrome on your existing paid plan; keep 'ask before acting' on for anything sensitive and let admins pin the site allowlist."
  },
  {
   "if": "You are building browser automation into your product",
   "then": "Prototype on Browser-use (free, MIT, any model); move to Browserbase + Stagehand when you need proxies, captcha solving, concurrency, and replay in production."
  },
  {
   "if": "The browser work is your own app \u2014 E2E tests, verifying what your coding agent just built",
   "then": "Playwright MCP. Free, deterministic, no stealth needed on sites you control, and every coding agent already speaks MCP."
  },
  {
   "if": "You automate other companies' logged-in sites at commercial scale",
   "then": "Get counsel before code. The Amazon v. Perplexity injunction fight (Ninth Circuit, argued Jun 11, 2026) is deciding whether authenticated agent access is a CFAA violation \u2014 and 4 of 7 agentic browsers failed a Jul 2026 same-origin security audit."
  }
 ],
 "field": [
  {
   "name": "Gemini in Chrome",
   "maker": "Google",
   "note": "Auto-browse agentic mode (Jan 2026, preview) inside the 3B-user browser, gated to AI Pro/Ultra; absorbed Project Mariner May 2026 \u2014 the biggest distribution threat to everyone above",
   "url": "https://gemini.google/overview/gemini-in-chrome/",
   "oss": false,
   "entry": "AI Pro sub (~$20/mo)",
   "status": "active"
  },
  {
   "name": "ChatGPT Atlas",
   "maker": "OpenAI",
   "note": "Standalone AI browser (Oct 21, 2025) retiring Aug 9, 2026 after ~9 months; agent capabilities folding back into ChatGPT and Codex",
   "url": "https://openai.com/index/introducing-chatgpt-atlas/",
   "oss": false,
   "entry": "\u2014",
   "status": "sunsetting"
  },
  {
   "name": "OpenAI Operator",
   "maker": "OpenAI",
   "note": "The Jan 2025 computer-use pioneer; deprecated into ChatGPT agent mode and shut down Aug 2025",
   "url": "https://openai.com/index/introducing-operator/",
   "oss": false,
   "entry": "\u2014",
   "status": "dead"
  },
  {
   "name": "Project Mariner",
   "maker": "Google DeepMind",
   "note": "Research browser agent (Dec 2024) folded into Gemini in Chrome, May 2026",
   "url": "https://deepmind.google/models/project-mariner/",
   "oss": false,
   "entry": "\u2014",
   "status": "renamed"
  },
  {
   "name": "Edge Copilot Mode",
   "maker": "Microsoft",
   "note": "Agentic browsing brand (Jul 2025) retired May 13, 2026 \u2014 features folded into core Edge, free",
   "url": "https://www.microsoft.com/en-us/edge",
   "oss": false,
   "entry": "free",
   "status": "renamed"
  },
  {
   "name": "Skyvern",
   "maker": "Skyvern (YC S23)",
   "note": "Vision+LLM workflow automation for enterprise forms \u2014 20k+ stars, 10M+ workflows run, SOC2/HIPAA, 500+ enterprise users",
   "url": "https://www.skyvern.com",
   "oss": true,
   "entry": "5k free credits/mo",
   "status": "active"
  },
  {
   "name": "Steel",
   "maker": "Steel (open source)",
   "note": "Open-source browser API for agents; self-hostable Browserbase alternative (~7k stars)",
   "url": "https://steel.dev",
   "oss": true,
   "entry": "free (self-host)",
   "status": "active"
  },
  {
   "name": "Hyperbrowser",
   "maker": "Hyperbrowser",
   "note": "Headless browser cloud purpose-built for AI agents; stealth and scale focus",
   "url": "https://hyperbrowser.ai",
   "oss": false,
   "entry": "usage-based",
   "status": "active"
  },
  {
   "name": "Airtop",
   "maker": "Airtop",
   "note": "Cloud browser automation with enterprise-compliance positioning",
   "url": "https://www.airtop.ai",
   "oss": false,
   "entry": "usage-based",
   "status": "active"
  },
  {
   "name": "Lightpanda",
   "maker": "Lightpanda",
   "note": "Open-source headless browser written in Zig \u2014 claims 11x faster, 9x less memory than Chrome for agent workloads",
   "url": "https://lightpanda.io",
   "oss": true,
   "entry": "free",
   "status": "active"
  },
  {
   "name": "Agent Browser",
   "maker": "Vercel Labs",
   "note": "Rust CLI browser for agents (~35k stars) \u2014 the fast-riser in the dev-tool lane",
   "url": "https://github.com/vercel-labs/agent-browser",
   "oss": true,
   "entry": "free",
   "status": "active"
  },
  {
   "name": "Amazon Nova Act",
   "maker": "AWS",
   "note": "Browser-action SDK + model (Mar 2025) with Playwright integration; enterprise/RPA lane",
   "url": "https://nova.amazon.com/act",
   "oss": false,
   "entry": "free preview + AWS usage",
   "status": "active"
  },
  {
   "name": "Manus Browser Operator",
   "maker": "Manus \u2192 Meta",
   "note": "Extension-based browser operator from the viral general agent \u2014 Manus acquired by Meta (2026, per manus.im); originally built on Browser-use",
   "url": "https://manus.im",
   "oss": false,
   "entry": "credit plans",
   "status": "acquired"
  },
  {
   "name": "Dia",
   "maker": "The Browser Company \u2192 Atlassian",
   "note": "AI browser acquired with TBC for $610M cash (Sep 2025); pivoting toward enterprise knowledge work; Arc left in maintenance",
   "url": "https://www.diabrowser.com",
   "oss": false,
   "entry": "free beta",
   "status": "acquired"
  },
  {
   "name": "Opera Neon",
   "maker": "Opera",
   "note": "Premium agentic browser (Neon Do/Make agents), $19.90/mo \u2014 a hard sell against free Comet",
   "url": "https://www.opera.com/neon",
   "oss": false,
   "entry": "$19.90/mo",
   "status": "active"
  },
  {
   "name": "Fellou",
   "maker": "Fellou AI",
   "note": "'Self-driving browser' with inspectable Deep Action workflows and open-source Eko framework",
   "url": "https://fellou.ai",
   "oss": false,
   "entry": "free download",
   "status": "active"
  },
  {
   "name": "Genspark AI Browser",
   "maker": "Genspark",
   "note": "Agentic browser running 169+ on-device open-weight models; $160M raised at $530M; Sep 2025 security audit flagged vulnerabilities",
   "url": "https://www.genspark.ai",
   "oss": false,
   "entry": "free tier",
   "status": "active"
  },
  {
   "name": "Firefox AI Window",
   "maker": "Mozilla",
   "note": "Opt-in, kill-switch-first AI browsing (Feb 2026); rated safest and least capable in the Jul 2026 UW audit",
   "url": "https://www.firefox.com",
   "oss": true,
   "entry": "free",
   "status": "active"
  },
  {
   "name": "Cloudflare Browser Run",
   "maker": "Cloudflare",
   "note": "Browser Rendering rebranded Apr 2026 with native WebMCP support; 120 concurrent browsers on Workers",
   "url": "https://developers.cloudflare.com/browser-rendering/",
   "oss": false,
   "entry": "Workers plans",
   "status": "renamed"
  },
  {
   "name": "MultiOn",
   "maker": "MultiOn \u2192 AGI Inc.",
   "note": "Early browser-agent API; multion.ai now redirects to theagi.company \u2014 pivoted to on-device phone/PC agents (AGI-0), browser product gone",
   "url": "https://theagi.company",
   "oss": false,
   "entry": "\u2014",
   "status": "renamed"
  },
  {
   "name": "Adept",
   "maker": "Adept \u2192 Amazon",
   "note": "The original 'AI that uses your software' startup ($415M raised); acqui-hired by Amazon Jun 2024 \u2014 the category's cautionary tale",
   "url": "https://www.adept.ai",
   "oss": false,
   "entry": "\u2014",
   "status": "acquired"
  }
 ],
 "signals": [
  {
   "fact": "The Great Consolidation: in ~60 days OpenAI announced Atlas's retirement (Jul 9 \u2192 dead Aug 9, 2026), Microsoft folded Copilot Mode into Edge (May 13), and Google folded Mariner into Gemini/Chrome (May) \u2014 agents are disappearing into surfaces people already use",
   "src": "https://nohacks.co/blog/agentic-browser-landscape-2026"
  },
  {
   "fact": "Manus \u2014 the viral agent built on Browser-use \u2014 is 'now part of Meta' (site banner, 2026): big tech is buying the layer, not just building it",
   "src": "https://manus.im/"
  },
  {
   "fact": "Amazon v. Perplexity is the first federal appellate test of whether AI agents may act in users' authenticated accounts: demand Nov 2025, preliminary injunction Mar 10, 2026 (stayed), Ninth Circuit argued Jun 11, 2026",
   "src": "https://nohacks.co/blog/agentic-browser-landscape-2026"
  },
  {
   "fact": "Security is the binding constraint: 4 of 7 agentic browsers let attackers bypass same-origin policy (Univ. of Washington, Jul 3, 2026); the safest (Firefox AI) was also the least capable",
   "src": "https://nohacks.co/blog/agentic-browser-landscape-2026"
  },
  {
   "fact": "Money kept flowing to infra: Browser-use $17M seed (Mar 2025), Browserbase $40M Series B at $300M (Jun 2025), Genspark $160M at $530M \u2014 while consumer AI browsers died or were acquired (Atlas, TBC/Dia $610M)",
   "src": "https://www.firecrawl.dev/blog/best-browser-agents"
  },
  {
   "fact": "Standards arriving: Playwright MCP at 6.3M weekly npm downloads (Aug 2026); WebMCP \u2014 sites exposing tools directly to agents \u2014 in Chrome 149 origin trial (Jun 2026), backed by Google and Microsoft via W3C",
   "src": "https://api.npmjs.org/downloads/point/last-week/@playwright/mcp"
  }
 ],
 "notes": "Ranking criteria: production-worthiness for a small AI-led company (safety controls, reliability, price) over raw autonomy demos; assisted browsing and programmatic automation are both in-scope jobs. Atlas retirement caveat: the Aug 9, 2026 shutdown (announced Jul 9) is confirmed by two independent secondary sources citing an OpenAI help-center article ('Evolving Atlas into ChatGPT'), but we could not fetch that primary page directly \u2014 re-verify. Conflict resolved: Anthropic's support doc still says Max-only beta while claude.com and the Dec 18, 2025 blog say all paid plans \u2014 we follow the newer product page. Amazon-injunction dates (Mar 10, 2026; Ninth Circuit Jun 11) are single-source (nohacks.co); the underlying dispute is primary-verified via Perplexity's Nov 2025 post. Browserbase's '50M sessions in 2025' is from a Firecrawl roundup, not Browserbase's own pages. Browser-use's Odysseys #1 is self-reported in its README; no neutral browser-agent leaderboard has category authority in Aug 2026 (WebVoyager numbers float 85\u201389% across vendors, all self-run). Adjacent elements: scraping/data-extraction APIs (Firecrawl) and computer-use model APIs belong with their model/data elements, not here; workflow tools that merely trigger browsers \u2192 Wf; agent loops you build yourself \u2192 Ag; MCP servers generally \u2192 Mc. Comet user counts: Perplexity has never published Comet MAU \u2014 treat any figure as modeled.",
 "sources": [
  "https://claude.com/claude-for-chrome",
  "https://claude.com/blog/claude-for-chrome",
  "https://support.claude.com/en/articles/12012173-getting-started-with-claude-for-chrome",
  "https://github.com/browser-use/browser-use",
  "https://browser-use.com/",
  "https://docs.browser-use.com/",
  "https://techcrunch.com/2025/03/23/browser-use-the-tool-making-it-easier-for-ai-agents-to-navigate-websites-raises-17m/",
  "https://www.browserbase.com/pricing",
  "https://www.browserbase.com/blog",
  "https://docs.browserbase.com/",
  "https://www.stagehand.dev/",
  "https://github.com/microsoft/playwright-mcp",
  "https://api.npmjs.org/downloads/point/last-week/@playwright/mcp",
  "https://www.perplexity.ai/comet",
  "https://en.wikipedia.org/wiki/Comet_(browser)",
  "https://www.perplexity.ai/hub/blog/bullying-is-not-innovation",
  "https://www.eesel.ai/blog/perplexity-comet-pricing",
  "https://gemini.google/overview/gemini-in-chrome/",
  "https://openai.com/index/introducing-chatgpt-atlas/",
  "https://nerova.ai/news/openai-retiring-atlas-browser-computer-use-moving-into-chatgpt-july-2026",
  "https://nohacks.co/blog/agentic-browser-landscape-2026",
  "https://www.firecrawl.dev/blog/best-browser-agents",
  "https://www.skyvern.com/",
  "https://manus.im/",
  "https://theagi.company/",
  "https://en.wikipedia.org/wiki/The_Browser_Company",
  "https://fellou.ai/"
 ],
 "element": {
  "number": 11,
  "name": "Browser Agent",
  "group": "Agents & Automation",
  "essential": false,
  "edition": "v2026.Q3",
  "revision": "r7",
  "license": "CC BY 4.0 \u2014 cite elems.ai",
  "url": "https://elems.ai/e/bu.html"
 }
}