Stagehand

the SDK for browser agents, built by @browserbase

Join the Discord:

[Stagehand @Stagehanddev]

Stagehand just crossed 1 Million weekly downloads, but we're not done yet.

New in our latest 3.5.0 release:
- native clipboard API
- screenshots in extract()
- better snapshots & local mode use

[Kyle Jeong @kylejeong]

https://t.co/Nuz1VIkktf

[Stagehand @Stagehanddev]

See full release notes here:

https://t.co/6fDFv5q1Rp

[Stagehand @Stagehanddev]

Stagehand 3.7 is out.

Here's what shipped last week:
- Control where your agent can navigate to domain allowlisting/denylisting
- CUA support for Gemini 3.5 Flash and the GPT-5.6 family (Sol, Luna, Terra)
- Various bug fixes & patches

pnpm add @browserbasehq/stagehand@latest

[Kyle Jeong @kylejeong]

GPT 5.6 Sol absolutely crushes on computer-use.

While barely losing to Fable in accuracy, it makes up in speed and cost, being ~3 times faster on average.

Grok 4.5 is also had a step change improvement, matching Gemini 3.5 flash in accuracy.

[Paul Klein IV @pk_iv]

If I could only follow one person on X, it would be @levie, CEO of Box.

We sat down to speak about why the REAL economy hasn't felt AI yet, the "model overhang" between capability and adoption, and what founders should actually build now that software is cheap.

00:00 - Cold open: agents use data like people do
01:53 - Is Redwood City the new AI epicenter?
05:18 - 90% of enterprise data is unstructured and was never automatable
07:46 - The model overhang: capability is outpacing adoption
10:35 - Why coding got the fastest AI takeoff
16:24 - Inside Box's agent-first rebuild
19:15 - Headless SaaS: agents will outnumber humans 100:1
25:41 - Model routing + why the applied layer wins, even bitter-lesson-pilled
27:45 - GTM is the new moat

Checkout Navigators, our new pod on Youtube, Spotify, and Apple Podcasts!

[Browserbase @browserbase]

If you didn’t see the Browserbase puppies, you missed out.

Last week we took over AIE, teaching the world’s best AI engineers how to build agents that reliably interact with the web.

We also launched Agents, a fully managed browser agent you can call with a single API.

[Kyle Jeong @kylejeong]

Anthropic is too good at Computer-Use.

Fable 5, Opus 4.8, and Sonnet 5 take the top 3 spots on our benchmarks using Stagehand.

[Browserbase @browserbase]

Computer use is finally reliable. It just needed a real harness. Now there is one.

Starting today, anyone can ship a SOTA browser agent, batteries included.

Introducing Browserbase Agents: one prompt and one API call is all you need to automate the whole web.

[Stagehand @Stagehanddev]

Try out the live demo:

https://t.co/qR5BCs7p91

[Stagehand @Stagehanddev]

Gemini 3.5 Flash + CUA is impressive: on our benchmarks, it performs similarly to frontier models while being on average ~80% cheaper.

It's a reliable model for powering browser agents that balance speed and accuracy on real-world tasks.

[Google AI Studio @GoogleAIStudio]

https://t.co/oFKtt2LToj

[Stagehand @Stagehanddev]

Docs here:

https://t.co/61kvJZvgsn

[Stagehand @Stagehanddev]

WebMCP support is live in Stagehand 3.6 🤘

Websites can now expose typed, first-party tools directly to your agent — Stagehand calls them instead of inferring from the DOM.

Plus: Azure OpenAI auth & smarter act caching.

[justine @justinemach_]

nobody's adopting WebMCP because site owners have to write and maintain them.

we taught browser agents to do it. webmcp-gen explores any page → generates the spec → stagehand injects and runs it.

here it is finding cheap flights on google flights:

[Browserbase @browserbase]

Excited to partner with Anthropic on Managed Agents,

The Browse CLI gives your agents reliable access to the web for advanced tasks.

Read the full documentation below.

[Kyle Jeong @kylejeong]

Claude Fable 5's computer use blows other foundation models out of the water.

On our custom task set it scores 7.3% higher than 4.8 while also being more cost efficient than sonnet 4.5.

Our hybrid agent built into Stagehand significantly outperformed the native CUA harness.

[Browserbase @browserbase]

Your agents shouldn't relearn how to use sites every time they visit.

https://t.co/vjMb1zPyEL is the web skills catalog for your agents. Since launching 2 weeks ago the community contributed over 300 skills and 35,000+ CLI downloads.

We're 2nd on Product Hunt, help us win product of the day!

[Stagehand @Stagehanddev]

@shreypandya massive

[Stagehand @Stagehanddev]

Stagehand just crossed 1 Million weekly downloads, but we're not done yet.

New in our latest 3.5.0 release:
- native clipboard API
- screenshots in extract()
- better snapshots & local mode use

[Stagehand @Stagehanddev]

Build browser agents with Stagehand:
https://t.co/fWAFijevgh

[Stagehand @Stagehanddev]

Checkout the page here:
https://t.co/5YJbn7eBdV

[Stagehand @Stagehanddev]

Introducing the new Stagehand Evals,

We've been working closely with labs evaluating their latest models on custom tasks that represent real, production use cases.

Excited to push forward the frontier and ensure our customers use the model best for them.