Category · Developer & infrastructure
Browser automation tools for AI agents
Servers that let an agent drive a real browser. Navigate, read accessibility trees, click, fill forms, capture traces and screenshots.
Best browser automation tools for AI agents, the shortlist284 head-to-head comparisons
Capability keys browser.control · browser.debug · All tools
letme.dev/browser.control picks the top-graded tool in this list and says how to call it direct (how letme works).
The same listing from the live API. Graded results come first, then the official MCP registry when no graded-only filter is set.
https://www.anchorterminal.com/api/v1/search
Filters
| Compare | # | Tool | Category | Grade | Score | Agent rating | p95 | Context | Price / x402 | Auth | Where | Details |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 19 | Browser UseBrowser Use · HTTP API | Browser | A | 81.5 | 3.0 (4) | n/a | n/a | $0.02 / browser-hr x402 | API key | Hosted and local | ||
|
Browser Use runs AI agents that operate a web browser. It has an MIT Python library and CLI, and a hosted cloud with an agent API, managed browsers over CDP and an MCP server. Top strength x402 host at x402.api.browser-use.com takes USDC on Base with no account or API key, and wallet top-ups start at $0.01 Top weakness The privacy policy says inputs are used to train Browser Use's AI models. Self-service zero data retention covers US projects only and doesn't say what model providers keep |
||||||||||||
| 98 | Chrome DevTools MCPGoogle (Chrome DevTools team) · MCP server | Browser | BB | 76.2 | 3.5 (8) | n/a | n/a | Free · OSS | None | Local | ||
|
Lets coding agents control and inspect a live Chrome instance. Top strength Performance traces, network inspection, heap snapshots and Lighthouse audits in one server Top weakness Usage statistics go to Google by default until you pass |
||||||||||||
| 137 | SeleniumSelenium project (Software Freedom Conservancy) · SDK + MCP | Browser | BB | 74.6 | 3.8 (4) | n/a | n/a | Free · OSS | None | Local | ||
|
Selenium is the open-source browser automation project behind the W3C WebDriver standard, with official bindings for Java, Python, C#, Ruby and JavaScript and a self-hosted Grid server that runs browsers on other machines. Top strength Official bindings for Java, Python, C#, Ruby and JavaScript, released together, 4.52.0 on 29 October 2026 Top weakness No MCP server from the project. The docs point to community servers and say they are third-party |
||||||||||||
| 143 | PuppeteerGoogle · SDK + MCP | Browser | BB | 74.4 | 3.3 (4) | n/a | n/a | Free · OSS | None | Local | ||
|
Puppeteer is Google's open-source Node.js library for controlling Chrome and Firefox over the DevTools Protocol and WebDriver BiDi. An agent writes and runs scripts that navigate, click, type, read pages and take screenshots. Top strength Twelve releases from 25.5.0 on 4 August to 25.16.0 on 22 October 2026, each with a changelog generated by release-please Top weakness No MCP server, CLI or llms.txt of its own. The docs point agents to the separate chrome-devtools-mcp project |
||||||||||||
| 179 | BrowserbaseBrowserbase · HTTP API | Browser | BB | 73.3 | 3.4 (8) | n/a | n/a | $0.12 / browser-hr x402 | API key | Hosted | ||
|
Hosted headless browsers for agents over CDP, plus Fetch and Search APIs and the Stagehand framework. Top strength x402 sessions at x402.browserbase.com with no account, $0.12 an hour, unused minutes refunded Top weakness The hosted MCP setup page passes the API key as |
||||||||||||
| 193 | SteelNen Labs, Inc. · HTTP API | Browser | BB | 73.1 | 2.0 (4) | n/a | n/a | $0.01 / call x402 | API key | Hosted and local | ||
|
Steel runs cloud browser sessions for AI agents, controlled through a REST API, a CDP WebSocket, SDKs, a CLI and an official MCP server. Its browser runtime, steel-browser, is open source and can be self-hosted. Top strength OpenAPI 3.0.3 document with 75 operations, llms.txt and a Markdown twin of every docs page Top weakness Two incidents of more than an hour hit session creation in the 90 days to 1 November 2026 (23 August, 1 October), per the status history read on 8 October |
||||||||||||
| 294 | Browserlessbrowserless.io, Inc. · HTTP API | Browser | BB | 70.4 | 3.5 (4) | n/a | n/a | $0.002 / credit | OAuth or key | Hosted and local | ||
|
Browserless runs hosted headless browsers that agents drive over CDP, Playwright, REST endpoints, BrowserQL or its official MCP server. The same software can be self-hosted from a Docker image. Top strength OpenAPI 3.0.0 document with 92 paths, llms.txt, llms-full.txt and Markdown twins of the docs Top weakness REST and WebSocket routes take the API token as |
||||||||||||
| 338 | TinyFishTinyFish · HTTP API | Browser | B | 69.3 | 2.8 (4) | n/a | n/a | $0.016 / call | OAuth or key | Hosted | ||
|
TinyFish is a hosted web agent platform from Tiny Fish, Inc. Agents call it through REST APIs for goal-driven automation, remote browser sessions over CDP, search, fetch and research, or through a remote MCP server with OAuth. Top strength OpenAPI 3.0 documents for seven APIs (Agent, Browser, Datasets, Fetch, Monitor, Research, Search), plus Top weakness The terms of 17 August 2026 allow Tiny Fish to train and improve its models on customer data. No opt-out was found in the reviewed pages |
||||||||||||
| 373 | HyperbrowserS2 Labs Inc. · HTTP API | Browser | B | 68.4 | 3.5 (4) | n/a | n/a | $0.001 / call x402 | OAuth or key | Hosted and local | ||
|
Hyperbrowser, from S2 Labs Inc., runs hosted Chromium sessions that agents control over CDP, WebDriver or computer actions, with Fetch, Search, Crawl, Extract and cloud sandbox APIs. Access is by API key, SDKs, a stdio MCP server or x402. Top strength x402 on /x402/session, /x402/web/fetch and /x402/web/search with USDC on Base or Solana and no account, confirmed by live 402 answers on 1 November 2026 Top weakness The official MCP server's last npm release is 1.0.25 of 13 April 2025, and it isn't in the official MCP registry |
||||||||||||
| 408 | agent-browserVercel (Vercel Labs) · MCP server | Browser | B | 67.9 | 2.8 (4) | n/a | n/a | Free · OSS | None | Local | ||
|
agent-browser is an open-source command-line tool from Vercel Labs that lets AI agents drive Chrome through accessibility-tree snapshots and element refs. It runs locally as a Rust binary and includes a stdio MCP server. Top strength Accessibility-tree snapshots give each element a ref such as Top weakness 0.39.3, 0.40.0 and 0.41.0 fixed allowlist overrides and saved tokens exposed to page scripts, with no advisory or SECURITY.md |
||||||||||||
| 410 | KernelKernel Technologies, Inc. · HTTP API | Browser | B | 67.9 | none | n/a | n/a | $0.06 / browser-hr | OAuth or key | Hosted | ||
|
Kernel (written KERNEL by the vendor) runs isolated cloud Chromium browsers for AI agents, driven by Playwright, CDP or computer controls, with stealth, proxies, managed logins and replays, through a REST API, SDKs, a CLI and remote MCP. Top strength Scoped API keys with resource and action policies and expiry since 16 October 2026 Top weakness Security incident disclosed 4 November 2026, with API keys and customer deployment secrets taken from logs |
||||||||||||
| 429 | Playwright MCPMicrosoft · MCP server | Browser | B | 67.5 | 3.0 (2) | n/a | n/a | Free · OSS | None | Local | ||
|
Browser automation MCP server built on Playwright that drives pages via structured accessibility snapshots rather than screenshots. Top strength Accessibility snapshots with Top weakness
|
||||||||||||
| 534 | PomeradoPomerado AI, Inc. · MCP server | Browser | B | 65.1 | 3.0 (4) | n/a | n/a | $50 / mo | OAuth or key | Hosted | ||
|
Pomerado builds and hosts deterministic tools that each do one task on one website, such as a flight search or adding to a cart, which agents call through a hosted MCP server or a REST API. Top strength 584 public read-only tools across 447 sites run with no account, key or card on the keyless MCP URL and the REST API, as listed on 2 November 2026 Top weakness First released on 5 October 2026, with a team of two per its YC profile |
||||||||||||
| 675 | NotteNotte Labs Inc. · HTTP API | Browser | B | 62.2 | 2.5 (4) | n/a | n/a | $0.01 / session-hr | OAuth or key | Hosted | ||
|
Notte, from Notte Labs Inc., runs hosted browser sessions with web agents, scraping, search, serverless browser functions, credential vaults and personas. Access is by REST API, Python and Node SDKs, a CLI or a hosted MCP server. Top strength OpenAPI 3.1 at api.notte.cc/openapi.json with 110 paths and 137 operations, plus llms.txt, Markdown twins of the docs and a keyless documentation MCP server Top weakness status.notte.cc records the Scrape API down for about 32 hours over 14 days in 90 days, including 20 hours from 23 October |
||||||||||||
| 686 | Anchor BrowserAnchor Forge Inc. · HTTP API | Browser | B | 62 | 3.3 (4) | n/a | n/a | $0.01 / call | OAuth or key | Hosted | ||
|
Anchor Browser runs cloud browser sessions for AI agents, controlled through a REST API, a CDP WebSocket, TypeScript and Python SDKs and hosted MCP servers. It adds managed logins with MFA, profiles, proxies, CAPTCHA solving and recordings. Top strength OpenAPI 3.1 document with 101 paths and 136 operations, llms.txt, llms-full.txt and a Markdown twin of every docs page Top weakness No service agreement or API terms found. anchorbrowser.io/terms is a website terms of use dated 16 March 2025 |
||||||||||||
| 826 | BxPlaywrightOrtus Solutions · SDK + MCP | Browser | C | 58.7 | none | n/a | n/a | Free · OSS | None | Local | ||
|
BxPlaywright (bx-playwright) is Ortus Solutions' Apache-2.0 BoxLang module for browser automation and testing on Playwright for Java, with browser tools for BoxLang AI agents and a command that starts Playwright's MCP server. Only a 1.0.0 snapshot is published. Top strength Ten compact in-process browser tools for BoxLang AI, each returning the URL, title and an accessibility snapshot with element refs Top weakness No stable release on 1 November 2026. ForgeBox has only 1.0.0-snapshot, though Ortus announced 1.0 as generally available on 13 October |
||||||||||||
| 828 | LightpandaSelecy SAS (Lightpanda) · MCP server | Browser | C | 58.7 | none | n/a | n/a | $19 / mo | None | Local | ||
|
Lightpanda is an open-source headless browser written in Zig, with the V8 JavaScript engine and no graphical rendering. Agents drive it on their own machine over CDP, WebDriver BiDi, a built-in MCP server or a CLI, or through Lightpanda Cloud. Top strength Version 1.0.2 of 23 October 2026, after six tagged releases since 16 August, with Docker, Homebrew, AUR, .deb, npm and pip install routes Top weakness Telemetry with the IP address and a persistent instance ID is on by default; tool and MCP client events in 1.0.2 aren't in the privacy policy's list |
||||||||||||
| 834 | SkyvernIkonomos Inc. (Skyvern) · HTTP API | Browser | C | 58.5 | 3.0 (4) | n/a | n/a | $29 / mo | OAuth or key | Hosted and local | ||
|
Skyvern is a browser agent that completes multi-step web workflows from natural-language goals using language models and computer vision. It runs as an AGPL-3.0 open-source server or a hosted cloud with a REST API and MCP server. Top strength Every MCP tool carries readOnlyHint, destructiveHint and openWorldHint annotations, 118 registrations in the source on 2 November 2026 Top weakness Every API key and OAuth token has full organisation authority. The docs state there are no read-only, per-endpoint or per-agent keys |
||||||||||||
| 921 | Kane CLITestMu AI (formerly LambdaTest) · HTTP API | Browser | C | 56.5 | none | n/a | n/a | $17 / mo | OAuth or key | Local | ||
|
Kane CLI is TestMu AI's (formerly LambdaTest) command-line testing agent. A person or coding agent gives a plain-English objective, and it drives Chrome or a mobile device and returns NDJSON results, with Exploration mapping a web app on HyperExecute. Top strength
Top weakness Exploration defaults to |
||||||||||||
| 934 | AirtopSwitchboard Visual Technologies, Inc. (dba Airtop) · HTTP API | Browser | C | 56.2 | none | n/a | n/a | $29 / mo | API key | Hosted | ||
|
Airtop runs cloud browser sessions that an agent controls through a REST API, Node and Python SDKs, natural-language page calls, or Puppeteer, Playwright and Selenium connections. The same company sells Agent Builder, a no-code product for scheduled web agents. Top strength OpenAPI 3.1 document with 39 operations, an llms.txt index and a Markdown twin of every docs page Top weakness No published rate for what a browser minute, a GB of proxy traffic or an AI call costs in credits |
||||||||||||
| 1041 | CurlShotCurlShot · MCP server | Browser | D | 53 | none | n/a | n/a | $19 / mo | API key | Hosted | ||
|
CurlShot captures web pages as images, PDFs or scrolling videos through a hosted MCP server and REST API. The REST API also renders supplied HTML and Markdown. Top strength Two MCP tools with typed arguments, a required URL and a format enum, documented in the agent guide Top weakness Capture access requires browser signup, verified email and a dashboard key |
||||||||||||
| 1222 | Functionize MCP serverFunctionize, Inc. · MCP server | Browser | D | 46.6 | 2.8 (4) | n/a | n/a | $20 / mo | OAuth | Hosted | ||
|
The Functionize MCP server is a hosted remote server from Functionize, Inc. that lets coding agents such as Claude Code, Cursor and GitHub Copilot build, run and diagnose browser tests of a live web app in Functionize Studio. Top strength OAuth only, with PKCE, dynamic client registration, a revocation endpoint and separate Top weakness Tool definitions, input schemas and annotations aren't published; they sit behind OAuth sign-in |
||||||||||||
| 1236 | BrowserActHongkong Ziniao Technology Co., Limited · HTTP API | Scrapers | D | 46 | none | n/a | n/a | $29 / mo | API key | Hosted | ||
|
BrowserAct turns a prompt into a reusable web scraper bot that runs in managed cloud browsers with proxies and CAPTCHA handling, called through a REST API, a remote MCP server or no-code tools, plus a local Agent CLI. Top strength OpenAPI 3.0.3 spec and llms.txt for the REST API Top weakness No status page, SLA, 429 guidance or idempotency keys |
||||||||||||
| shut | Puppeteer (archived MCP reference server)Shut down 2025-05-29 MCP project (archived reference servers) · MCP server | Browser | F | 34.1 | 1.0 (2) | n/a | n/a | Free · OSS | None | Local | ||
|
The original browser-automation reference server, driving Chrome through Puppeteer with screenshots and JavaScript evaluation. Top strength Seven tools in about 700 tokens, every input with a Top weakness Archived on 29 May 2025 with no security guarantees |
||||||||||||
Nothing matches these filters. .
p95 latency and context cost come from our probes, which haven't run yet, so those columns start hidden. Grades run from AA to F, and agent-ready means BB or better. Filters, sorting and export run in your browser; the table is complete without JavaScript.
Indexed, not reviewed (31)
Listings sorted into this category from public catalogues (the official MCP registry, APIs.guru, the x402 Bazaar and OpenRouter), with facts and our own checks but no score, grade or rank. How the index works.
| Listing | Kind | What it does | Why it's here |
|---|---|---|---|
| Actionbook actionbook.app | MCP server | Let your AI agent operate any website through a real, logged-in browser session. | vendor's own, widely used |
| AgentUtility Browser Workflow agentutility.ai | MCP server | MCP server for the @agentutility browser-workflow cluster — pay-per-call x402 tools, no API keys… | vendor's own |
| Aginx Browser aginx.net | MCP server | Stealth web browser for agents: search, fetch, click, download and type in persistent MCP sessions. | vendor's own |
| AppContext appcontext.dev | MCP server | Live iOS Simulator and browser screenshots, logs, and a personal MCP hub for AI coding agents. | vendor's own |
| browsa browsa.io | MCP server | Cloud browsers for AI agents: run any web task in plain English on Phantom, Browsa's own browser. | vendor's own |
| Browz gobrowz.com | MCP server | Isolated, fingerprint-protected browser profiles for AI agents: one per task, with its own proxy | vendor's own, widely used |
| buy usebuy.ai | MCP server | Pay per request for APIs, compute, browsers, RPC and LLM inference with stablecoins on Celo. | vendor's own |
| Call Grok Bot zens.ai | MCP server | Consult existing Grok Bots from Codex and Claude Code using official Cursor browser login. | vendor's own |
| Endform endform.dev | MCP server | Find failing, flaky, and slow Playwright tests on Endform and keep your suite healthy as it grows. | vendor's own, widely used |
| foura.ai MCP server foura.ai | MCP server | Reliable web access for AI agents: smart HTTP, rotating proxies, and full-browser rendering. | vendor's own |
| ghostlight sylin.org | MCP server | Visible local browser automation in signed-in Chromium, with optional policy and audit. | vendor's own |
| GRID Agent Tools grid.is | MCP server | Load, read, edit, and recalculate .xlsx workbooks headlessly — no Excel required. | vendor's own |
| GrowBlocks Dev Kit growblocks.xyz | MCP server | Build, play-test with screenshots, and ship three.js browser games on GrowBlocks. | vendor's own |
| openbrowser-ai openbrowser.me | MCP server | AI browser automation. Write async Python to navigate, click, type, and extract data. | vendor's own |
| Qiksy Bridge qiksy.app | MCP server | Browser MCP for the Chrome you already have open, via the free Qiksy Bridge extension. | vendor's own |
| QuickPeek quickpeek.co | MCP server | Turn a URL into a narrated demo video: real browser walkthrough, AI voiceover, captions burned in. | vendor's own |
| reflex reflexmcp.com | MCP server | Browser automation for AI agents: compressed page snapshots, batched actions, fewer tokens. | vendor's own |
| secondlayer secondlayer.tools | MCP server | Stacks chain data: contracts, events, transfers, subgraph scaffolds. Hosted with browser sign-in. | vendor's own, widely used |
| Shardflux shardflux.dev | MCP server | Persistent cloud workspaces for AI agents: run commands, edit files, use git and a browser. | vendor's own |
| Shiplight shiplight.ai | MCP server | An MCP server that provides browser automation | vendor's own |
| slickfast.com MCP server slickfast.com | MCP server | Charts and dashboards as images for AI agents. No browser, same data same bytes. Local and free. | vendor's own |
| Squally squally.dev | MCP server | Query Playwright test analytics from AI assistants: flaky tests, failures, CI time lost to retries. | vendor's own |
| TabRunner tabrunner.app | MCP server | Give browser tasks to TabRunner, an agent in your own Chrome that uses your logins. | vendor's own, widely used |
| Tada: automated QA for AI-written code taada.dev | MCP server | Your coding agent asks Tada to work through a feature in a real browser and gets back a verdict. | vendor's own |
| TurnSignal for Playwright turnsignal.ai | MCP server | Read-only Playwright CI results for agents: new vs known failures, flaky tests, test history. | vendor's own |
| Usercase usercase.io | MCP server | Ask real users on your site and run virtual users in a real browser, then fix what they hit. | vendor's own, widely used |
| walkerOS Source Browser walkeros.io | MCP server | Generate, parse, and validate walkerOS data-elb HTML tagging attributes | vendor's own, widely used |
| Weavetab weavetab.dev | MCP server | Open automation engine for AI agents. Control live browsers and desktop apps locally. | vendor's own |
| Web Browser (kogiQA) atagon.com | MCP server | This web browser has been designed to help your agent debug and develop complex web applications. | vendor's own, widely used |
| Wield wield.run | MCP server | Websites as tools for AI: hosted for public reads, local in your own browser for logins. | vendor's own |
| Yab, Yet Another Browser yetanotherbrowser.com | MCP server | Your own signed-in Mac browser for any agent: open, read and act in tabs. Local and free. | vendor's own |
How the ranking works
Every tool is scored 0 to 100 across 9 weighted categories, minus deductions for negative events, and given a grade from AA to F. Reliability, latency and context cost are measured by probes and by traffic through letme. Schema quality, security and payments are assessed against a published checklist. Task success comes from a fixed task suite. Payments, including x402, are one category of the nine. Read the methodology.
For companies
Do agents find, use and choose your tools?
An agent-readiness audit runs our probes, task suite and eight reviewer agents against your public and internal tools, and comes back with a scorecard, the transcripts of what failed, and a fix list in priority order. From $2,500, re-run included. We never take payment to move a rank. We do help companies earn one.