Add one line of code to run any browser automation in the cloud.
Kernel takes your browser automation code, puts it in a virtual machine alongside a browser, and executes it in the cloud. Say goodbye to managing Kubernetes, Docker containers, and AWS acronyms.
Bring your own code
Build automations using browser frameworks like Playwright or write your own Computer Use agentic loop
Debuggability at your fingertips
Use our Live View to debug sessions and build human-in-the-loop workflows
Insanely fast spin-up
Launch web agents in 20 milliseconds or less. Our unikernel-based runtime is built for speedy cold starts
Restore & reuse
Return to an agent's browser hours or days later. Reuse auth sessions, cookies, browser history, and more
Invoke your agent
Schedule your agent to run on a cron, invoke it on-demand, or serve it as an HTTPS endpoint
Never pay for idle time
Kernel browsers automatically pause and resume based on network activity. You only pay for compute you actually use
Open source
We're building the next generation of AI infrastructure. Help steer our roadmap! Accepting external contributors
Step 1
Join our waitlist
Join our private beta for early access.
Step 2
Deploy your app
Install @onkernel/cli, our command line interface, and run kernel deploy where you've built your app. It's as simple as that.
Step 3
Invoke on-demand
Run your actions on a schedule or invoke them from other parts of your app. We handle all the infrastructure.
Free
+ usage costs
- Browser pause & resume
- Logs & replays
- Single browser (no concurrency)
- Community support
$200 / mo
+ usage costs
- File I/O
- Cron scheduling
- User auth & secrets management
- Concurrent browsers
$450 / mo
+ usage costs
- BYO images & storage mounts
- Configurable browser extensions
- SOC2 & HIPAA compliance
- Advanced support & SLAs