Home / Directory / Agents & orchestration / AI browser agents & web automation / Kernel
Kernel
Cloud browser infrastructure from San Francisco that runs Chromium on unikernels, built for AI agents, web automation, and browser testing.
Engineering teams running browser agents or end-to-end test suites that need fast cold starts, logged-in sessions, and replays they can debug from.
Ops teams that want a no-code tool to automate a website, or buyers who need a long track record from a browser vendor.
Verdict
Kernel is a browser infrastructure company in San Francisco, started in 2025 by Catherine Jue and Rafael Garcia. Developers use it to spin up cloud Chromium browsers for AI agents, scraping, and automated tests, with managed logins, stealth settings, live view, and video replays of each session. It stands out because it runs each browser on a unikernel, so a new browser is ready in about 30 milliseconds and idle browsers cost nothing, which matters when an agent or test suite opens browsers by the thousand.
Score Breakdown
How Kernel scores in the categories that matter to its buyers.
Pricing
Kernel publishes monthly plans that include usage credits, then bills browser time by the GB-second. Idle time and proxies are not billed.
| Plan | Price | What stands out |
|---|---|---|
| Free | $0 / month | $5 of usage credits, 5 concurrent browsers, 3 managed auth connections, 1-day replays |
| Hobbyist | $30 / month | $10 of credits, 10 concurrent browsers, unlimited proxies, 7-day replays |
| Start-up | $200 / month | $50 of credits, 150 concurrent browsers, bring your own proxies, GPU browsers, 30-day replays |
| Enterprise | Custom | 99.9% SLA, SSO, zero data retention, white-labeling; SOC 2 and HIPAA |
The Field at a Glance
How Kernel compares with other AI browser agents & web automation vendors we reviewed, by Overall Score and relative typical engagement cost.
Kernel scores 7.6 in AI browser agents & web automation, level with Browserbase (7.6), and ahead of Anchor Browser (7.4) and Skyvern (7.4). Relative cost lands mid-band for this subcategory.
Use-case matrix
| Use case | Fit | Notes |
|---|---|---|
| Browser agents at high volume | Strong | Unikernel browsers start in about 30 ms and scale to many at once. |
| End-to-end test suites | Strong | Stably moved its customers' test runs onto Kernel for speed and stability. |
| Logged-in sessions on real sites | Strong | Managed auth and stealth settings are built in. |
| Debugging agent runs | Strong | Live view and mp4 replays of each session. |
| No-code website automation | Weak | Kernel is infrastructure for developers, not a point-and-click tool. |
| Long vendor track record | Mixed | The company started in 2025, though it reports more than 11,000 teams. |
Who it’s for
Good fit
- AI agent startups that open many browsers
- QA platforms running customer test suites
- Teams replacing self-hosted headless Chrome
Poor fit
- Ops teams without engineers
- Buyers who need a decade-old vendor
- Teams that only automate a few pages a week
Review Excerpts
Below are excerpts from public reviews. Our team scoured public reviews, forums, and chat rooms to get a balanced view of customers' experience with this company. Paid reviews and pay-for-play sites such as Clutch were excluded.
“we wanted a fast end-to-end experience for our customers, but we really didn’t want to be in the browser infrastructure business. it’s really complicated to operate browser infrastructure. we have enough problems to solve.”
“we evaluated many options, and kernel is the only one that could deliver the performance all of our customers' test suites required.”
“it's been great not having to be an expert on managing headful chromium instances. we love kernel.”
“The free plan only keeps replays for one day. I went to debug a flaky run from Saturday and the recording was already gone on Monday, so we moved up to the $30 Hobbyist plan to get a week of replays.”
Methodology
This page is an independent evaluation of Kernel for buyers comparing options in ai browser agents & web automation. AI Industry Reviews accepts no sponsorships, advertising, or pay-for-placement fees. Kernel did not pay for this review.
What we scored
The headline number is an Overall Score on a 0-10 scale. Eight criteria sit under it in two groups.
- Buyer outcomes (for ai browser agents & web automation)
- Cold start speed & scale
- Auth, stealth & replays
- Developer experience
- Maturity & track record
- Company & commercial
- Innovation & product leadership
- Project management & communication
- Pricing
- Contract fairness
Pricing measures whether the price looks fair for the value delivered, including packaging and renewal friction that show up in real buying cycles.
Score Composition
| Input | Weight | What it covers |
|---|---|---|
| Reviews | 40% | A proprietary read of what practitioners say about likes, complaints, and day-to-day use, including public review sites, forums, and private chat rooms. Paid reviews and pay-for-play sites such as Clutch are out of scope. |
| Product | 35% | Hands-on look at screens and workflows. |
| Pricing | 15% | Whether the price looks fair for what you get. |
| Docs & training | 10% | Docs, tutorials, and training material. |
How we balanced the evidence
The Overall Score is the simple average of the eight criteria. Recommendation language follows that score and the fit pattern described above.
Scope
Kernel is graded here as ai browser agents & web automation. Criteria scores can move as more review volume and product checks are added.