// from the team
Product updates, engineering notes, and practical guides for building browser agents.
1–9 of 9 posts
Stagehand 4.4 adds browser.disconnect() for reattaching to running browsers, an extension watchdog that cleans up after crashed clients, working keepAlive on local browsers, and LangChain Deep Agents browser tools built on Stagehand.
Learn how to reliably extract typed data from any website using only 50 lines of code.
Use Stagehand with the official OpenAI and Anthropic SDKs, try new model IDs, and build workflows with page condition waits and flexible browser connections.
Codex is a powerful AI assistant that can use a browser to accomplish complex tasks. We benchmarked Codex using Playwright vs. Stagehand to perform browser tasks, and found that Stagehand's in-browser runtime architecture makes your overall Codex tasks twice faster.
Stagehand 4.2 adds WebMCP tools that change with the page, native PDF exports, timeouts for locator operations and snapshots, and more reliable iframe and shadow DOM interactions.
We turned 10 web pages into six LLM context formats, from raw HTML and Markdown to accessibility trees and screenshots, and measured token cost, read accuracy, and act accuracy. Fewer tokens rarely meant better answers.
Playwright MCP's accessibility snapshots pile up in context turn after turn. We benchmarked where the tokens go, how to trim them, and how Stagehand compares.
Stagehand 4.1 adds first-party Browserbase Search and Fetch, improved WebMCP iframe support, and clearer close behavior.
Meet Stagehand, the open-source SDK for browser agents, and see how browser control and AI methods work together to check an order’s shipping status.
