Anthropic released a new Browser Use tool Thursday that gives Claude a structured view of web pages, letting the AI model target elements by reference instead of guessing at screen coordinates. The tool relies on the page’s accessibility tree, which describes links, buttons, and text boxes in a machine-readable format. This is part of a wider release that also puts Computer Use, the Skills API, and the Files API into general availability. Developers can access the browser tool through the Claude API using browser_toolset_20260801.
Page references replace coordinates
The shift means Claude no longer has to estimate where a button sits on a viewport image and fire off commands like x: 640, y: 320. Instead, it can receive a reference such as ref_3 tied to a specific element and use that when it wants to act. When Claude calls read_page, the developer’s executor returns a text representation of the accessibility tree, with elements tagged for later use.
Computer Use still operates across an entire desktop by analyzing screenshots and sending mouse coordinates and keyboard commands. Browser Use works inside the browser itself, where it can lean on page structure that would be hard to recover reliably from pixels alone.
Related: Six Capabilities to Secure Autonomous AI Agents
There is a catch though. If the tab handles to a new page or the page changes significantly, a reference that pointed to a button a moment ago may no longer be valid. The API won’t catch that on its own. The executor has to recognize when the reference no longer matches the underlying element, reject the action, and have Claude read the page again before continuing.
Similar tools already exist
Playwright, for example, can represent a page as an ARIA snapshot and locate elements by role rather than coordinates. Microsoft’s Playwright MCP server already exposes structured accessibility snapshots with references a model can use to identify elements. The concepts line up closely with Browser Use, but the protocols do not: Playwright MCP speaks MCP, while Anthropic’s tool uses its own client-toolset protocol. Developers would still need an adapter that translates Claude’s requests into Playwright actions and returns the results in the format Claude expects.
Puppeteer offers many of the same building blocks, exposing the browser’s accessibility tree and providing APIs for controlling Chrome and Firefox. A developer could use those APIs for navigation or page reads, then maintain Anthropic’s reference mappings on top.
To make things slightly confusing, an unrelated open-source project also called Browser Use runs AI browser agents against Chromium through the Chrome DevTools Protocol. Despite the shared name, it has no connection to Anthropic’s tool and comes with its own agent loop and browser abstractions, so connecting the two would still require integration work.
Related: VMware Users Face Chaos Under Broadcom
Multiple actions in a single turn
Anthropic is also cutting down the back-and-forth between Claude and the browser by allowing multiple actions to be requested in one model turn. Actions can now arrive together as several tool_use blocks. The application executes them in order and sends the results back together, avoiding another model call between every click and keystroke. That can lower latency and costs, particularly as workflows scale from a handful of interactions to dozens or hundreds.
This matters more as browser tasks get longer. Cheaper models alone won’t solve the token cost problem in agentic workflows, so cutting unnecessary model calls is another way to reduce expenses. If Claude has to return to the model after every click or keystroke, a long task can quickly rack up model calls. Batching cuts out some of that back-and-forth, but the browser still has to carry out actions in order because each one depends on what happened before it. If Claude asks to click a button, fill in a field, and submit a form, the executor cannot simply move on if that first click fails, because everything that follows is based on a page state Claude never reached.
Browser Use is currently limited to the Claude API and is not available inside Claude Managed Agents. Adding it to a Messages API request exposes 27 browser operations by default. Claude can decide which operations it wants to use, but Anthropic does not execute them. The application has to translate each request into an action inside its own browser environment, preserve the session between turns, and return enough information for Claude to understand what happened.
Loading all of those operations has a token cost. Anthropic’s pricing documentation says the default Browser Use toolset adds roughly 6,600 input tokens to a request, before counting screenshots, accessibility trees, and other results sent back to Claude. Developers can turn off operations they don’t need to reduce that overhead.
Related: Extreme Networks Sees Strong Market Growth Ahead
Security and hosting considerations
Browser sessions, along with their downloads and uploaded files, stay in the developer’s environment. That differs from some of the other tools Anthropic announced Thursday — Skills uploaded through the Skills API can run inside Anthropic’s code execution sandbox, while the Files API stores documents that can be reused by ID.
Claude can still encounter a prompt injection in web content or be redirected to an unexpected location, which is why Anthropic recommends running the browser in an isolated container or virtual machine with minimal access. JavaScript and file uploads should remain disabled unless needed, since code generated by Claude runs with the page’s privileges and can reach data or make requests available to that page.
Batching makes approval a little trickier because several actions can arrive at once, and a routine click at the beginning of a sequence could eventually lead to something that requires the user’s permission. The executor has to check actions as they happen and stop for approval when needed.
