Browserbase

Your customers shouldn't have to trust agents blindly. With the Session Replay API, embed videos in your product and let customers watch the agent work within seconds of the session ending. The same replays we serve in our dashboard, now in yours.

@KenzieMac_Dev @resend Browserbase 🤝Resend

online shopping would be so much more fun if everything was always in your price range so i built that. chrome extension powered by @browserbase + @resend, full breakdown👇

Check out the docs: browserbase.run/MFuk3WE

Your customers shouldn't have to trust agents blindly. With the Session Replay API, embed videos in your product and let customers watch the agent work within seconds of the session ending. The same replays we serve in our dashboard, now in yours.

Turn any website into an API with /browser-to-api. This skill analyzes network activity, CDP logs, and website behavior to generate a custom OpenAPI spec. Watch Codex one-shot a fully documented OpenTable API client from a single prompt 👀

Save tokens and reduce context bloat with Stagehand 3.4.0. You can now use ignoreSelectors in extract() and observe() to keep noisy page regions out of model context. Plus: better agent defaults, new CUA models, and stronger frame handling.

Our refreshed Downloads API lets you download files from any page on the internet, perfect for data extraction workflows. You can easily fetch them from our s3, or (if you're a scale customer) store them in your own storage buckets.

Every AWS Lambda invocation runs in a full VM that boots in under 125ms. Firecracker is the ~50,000 line Rust binary that makes that possible. I wrote an interactive blog about it, with components you can play with.

Containers are fast. VMs are safe. Everyone building agent infra needs both. That's not just an opinion, it's the problem Firecracker was built to solve. I wrote about how it works with interactive components.

did you also forget to get your mom something for mother's day? i gotchu type her birth year + hometown → browserbase scrapes charts from her teenage years → claude picks the most iconic songs → spotify playlist, auto-created one command. she cries. you're forgiven.

OpenClaw can use Autobrowse to create and iteratively improve a Skill for any workflow. In this Craigslist extraction example we found a 68% speed up and 91% cost savings in 5 iterations. Halfway through, the agent found an exposed endpoint to speed up page navigation.

Read the full case study here: browserbase.run/shopvision

ShopVision runs 1.3M browser sessions a month and tracks 150k e-commerce sites daily. They cut their data pipeline from 13 hours to 3 by building a system that writes, tests, and fixes its own scripts. No extra engineers needed, just Browserbase & Stagehand getting it done.

@magicareai This is true

if you're not using @browserbase you're ngmi

Next up, a panel discussion with @pk_iv (@browserbase), @abhiaiyer (@mastra), @bhavz (Drata), and Rob Ferguson (@FireworksAI_HQ). Moderated by Auth0's @yenkel, the panelists share how they are enabling and using agents at work.

Hermes Agent can use Autobrowse to make better browser skills. In this HN example, 2 iterations take us from: 102 seconds -> 35 23 turns -> 8 $1.46 -> $0.28 Instead of clicking step by step it decides to eval JS directly on the page and save that workflow to a skill.

@KenzieMac_Dev Don't stress about infra, use Browserbase instead

I rage-built an agent to find events with free food. Now it runs every Monday without me touching it. Here's how I deployed it while managing zero infra. 👇

Read the post in our blog as well: browserbase.com/blog/autobrowse

Browser agents have an amnesia problem. Every run, they rediscover the same site from scratch. You pay the full exploration tax 100 times over and still have nothing to hand a teammate. We built Autobrowse to fix that, read more about how it works.