tools
Stagehand
Open-source SDK for building browser agents that act, observe, extract data, and run workflows using natural-language instructions.
In inglese
Stagehand helps developers automate real websites with browser actions such as clicking, filling forms, navigating, observing actionable elements, and extracting structured data. Its AI resolves instructions at runtime instead of relying only on fixed selectors.
It works locally with Chromium and can connect to Browserbase for cloud browsers, session replay, observability, captcha solving, and deployment. Stagehand itself is open source; Browserbase infrastructure, model usage, and related services may cost extra.
Features
- act() performs browser actions from plain-English instructions
- extract() returns structured page data with Zod schema validation
- observe() identifies actionable elements before an action
- agent() runs multi-step browser workflows autonomously
- Works locally with Chromium without requiring Browserbase
- Connects to Browserbase cloud browsers for production deployment
- Open source under the MIT license
- Supports TypeScript, Python, and Go SDKs
Use cases
- Automate multi-step workflows across changing websites
- Extract structured information from pages without a site API
- Test web applications with actions that adapt to DOM changes
- Navigate login-protected portals and fill forms
- Deploy browser agents on Browserbase cloud infrastructure
Pros
Cons
Latest updates
- Configurable caching in Stagehand (v4)
Stagehand v4 caches act(), observe(), and extract() results on Browserbase's servers. Now you control it.
- Introducing Stagehand v4: the SDK for browser agents (v4)
Improved context management, self-healing actions, and iframe support; Stagehand now lives in the browser as an extension.
- Cut inference costs 30–40% with Model Router
Set model: "auto", and Browserbase selects the model for each call server-side, balancing accuracy, speed, and cost automatically.
- Domain policies in Stagehand 3.7 (3.7.0)
Use setDomainPolicy() to allowlist or blocklist the domains your agent can visit.
Capabilities
- Choice of models — “Stagehand supports all major model providers — including OpenAI, Anthropic, and Google Gemini — via the Vercel AI SDK.” source
- API — “Perform browser actions from a plain-English instruction. Click, fill, navigate, scroll.” source
- Official SDKs — “Stagehand is an open source SDK that uses AI to make your browser agents resilient, readable, and production-ready.” source
- Builds agents and workflows — “Use agent when you need multi-step workflows executed autonomously, like navigating a complex checkout flow.” source
- Traces and evaluates — “session replay, prompt observability” source
Get it
Security
Pricing
- Starting price
- $20/mo
- Prices checked
- 2026-09-25
Free Plan
Free
- 3 concurrent browsers
- 1 browser hour
- 3 Agent runs included
- 1,000 Search calls
- 1,000 Fetch calls
- 15 minutes /session
Developer Plan
- $20/mo
- then $0.12/browser hr
- Then $12/GB
- 25 concurrent browsers
- 100 browser hours
- 15 Agent runs included
- 1,000 Search calls
- 1,000 Fetch calls
- 1 GB of proxies
Startup Plan
- $99/mo
- then $0.10/browser hr
- then $10/GB
- 100 concurrent browsers
- 500 browser hours
- 50 Agent runs included
- 1,000 Search calls
- 10,000 Fetch calls
- 5 GB of proxies
Scale Plan
Price on request
- 250+ concurrent browsers
- 500+ browser hours
- Custom Agent runs
- 10,000+ Search calls; usage-based
- 10,000+ Fetch calls; usage-based
- HIPAA (BAA) & DPA; SSO available