KERNEL

Introducing Browser Pools — instant browsers with the logins, cookies, and extensions your agents depend on. Designed to make using Kernel even faster.

@PriyankaPhatak congrats on the launch, we've loved partnering with you

Yesterday at Code with Claude: Tokyo, we launched support for 2 new exciting features for Claude Managed Agents. Scheduled Deployments: We introduce deployments as a new resource that lets to trigger sessions on a schedule. A really good example of this is something I personally use is getting the agent to scan and triage my teams linear backlog and tag assignees when the due date is past or if a ticket status hasn't changed in 2+ weeks which allows for me to maintain our linear board. Environment Variable in Vaults: This allows you to add environment variables in your Vaults like an API Key which unlocks the ability for you to use cli's with Claude Managed Agents. Here is an awesome cookbook from our friends over at @usekernel on how you can get Browser Use with CMA using their CLI github.com/kernel/claude-…. You can also use other cli's like sentry, github etc. Try out these features and give us feedback!

@AtomMccree @AnthropicAI @claudeai nice, if you've been running it locally for a month, the next unlock is parallel. kernel spins browsers up in the cloud so your agent runs across a bunch of them at once, even while you're away. worth a try

@usekernel @AnthropicAI @claudeai I been doing this with Claude Code + Claude Chrome for a month.

we've partnered with @anthropicai to give @claudeai managed agents browser capabilities for the first time. now, managed agents can navigate and interact with the web to complete long horizon, scheduled, and complex tasks — without managing infra. x.com/claudeai/statu…

@claudeai we love this launch and are excited for our integration that gives managed agents access to the live web get started with our cookbook: github.com/kernel/claude-…

New from Code with Claude Tokyo: scheduled deployments and environment variables in vaults are in public beta in Claude Managed Agents, and dynamic workflows in Claude Code are generally available. Agents now run on a schedule, use your tools securely, and take on bigger jobs.

@AnthropicAI @claudeai give your @claudeai managed agent a browser with our cookbook: github.com/kernel/claude-…

@AnthropicAI @claudeai until today, @claudeai managed agents were siloed to first-party integrations. with kernel, they now have browser access and managed auth across any authenticated service — without the agent ever seeing or touching user credentials. our blog shares more: kernel.sh/blog/claude-ma…

@stripe @Wix @laravelphp @HeyGen @prisma @automattic @p0 @ExaAILabs @Base44 @ClickHouseDB @supermemory @blaxelAI @getmetronome @AgentPhoneHQ @e2b @postalform we're thrilled to be a part of this launch

New providers now available on Stripe Projects: @Wix @laravelphp @HeyGen @prisma @automattic @p0 @ExaAILabs @Base44 @ClickHouseDB @supermemory  @usekernel @blaxelai @getmetronome @AgentPhoneHQ @e2b @postalform

@stripe to learn more about our partnership with @stripe, see: kernel.sh/blog/stripe-pr…

@stripe use the @stripe cli to spin up kernel browsers with 2 loc. stripe projects allows your agent to connect to kernel, choose a plan, and provision api keys without ever using the dashboard. get started here: kernel.sh/docs/integrati…

kernel is now integrated with @stripe projects. this partnership moves us closer to a world where an agent's stack is simply invokable, composable code. x.com/stripe/status/…

@erulabs we're hiring :)

It’s a rare day I say to myself “why in the hell don’t I work for these folks” and today is one of those days

agents are changing infra requirements. our co-founder and ceo, @juecd__ spoke at the replay conference by @temporalio to share lessons learned from giving agents access to the internet at human speed youtube.com/watch?v=ZqA3R1…

@strakedata love to hear it

@usekernel Very cool article! CoW forking and UFFD for firecracker sound both very interesting. Definitely will look into these and try some benchmarks

@raziborsky yes, we ran our firecracker microvms on a unikernel: news.ycombinator.com/item?id=437051… ultimately we wanted more control and performance so we added optimizations

@usekernel Did you look at unikernels as well? How did you make the decision to go with this?

the trick that makes @usekernel fast: we pay the full 5-10s of starting chromium once, snapshot ram to nvme, and resume in ~30ms when a request lands. from the outside it looks like a giant warm pool. under the hood it's mostly cold disk. wrote up how we got here.

@stably_ai read their full story. kernel.sh/customers/stab…

.@stably_ai uses kernel to build an ai-powered end-to-end testing platform. by switching to kernel, they saw 333x faster cold starts and eliminated infra-caused test flakiness. "we evaluated many options, and kernel is the only one that could deliver the performance all of our customers' test suites required." - jinjing liang, co-founder & ceo, stably

@useComputed @rfgarcia 💚💚

@rfgarcia @usekernel Having dealt with browser automation a ton in the past. This is unbelievably smart