Browser QA for teams shipping with AI

AI builds fast. Kery tests at that speed.

Kery runs autonomous browser QA for teams shipping with AI: real flows, real evidence, reproducible bugs.

Install locally Join Cloud waitlist
kery run / checkout
yourapp.example.com yourapp.example.com/cart live browser
Cart
Mango Hoodie $42.00
Remove
Agent Sticker Pack $8.00
Remove
Debug Tote $18.00
Remove
kery prompt ready
Describe a flow
› Hey kery, test my checkout flow
Collected bugs
  1. #001
  2. #002
  3. #003
  4. #004

Works with the tools your team already uses to build

Claude Code Codex Cursor Any MCP client
Run Kery locally

One command starts the open-source browser testing stack.

  • Open-source engine you can inspect, run, and extend locally.
  • Runs locally with Docker.
  • Installs MCP into Cursor, Claude Code, and Codex CLI.
  • Point Kery at your app running locally and start testing it.
  • Supports the following providers.
OpenAI Anthropic Google Gemini OpenRouter
View GitHub README See how it works
kery setup
kery cloud ci connected
GitHub Pull request
PR #184 · Add checkout promo 4 commits into main
Open
checkout.tsx cart.tsx promotions.ts

Intent detected: verify the promo path through cart and checkout.

Vercel Preview deployment
Deploying preview Deployment ready Preview for PR #184
Preview URL checkout-promo.vercel.app
GitHub commit 7f32c9a ready for browser tests
Kery Cloud targeted browser QA
Live target Running tests on checkout-promo.vercel.app
  1. ✓Open preview
  2. ✓Read PR intent
  3. !Test cart flow4 issues
  4. ✓Test checkout promo
GitHub Timeline update
Kery Cloud reported QA results PR #184 timeline update
Now
Commented with screenshots and repro steps
2 issues created
cart flow checkout promo repro attached
Kery Cloud

Cloud QA for teams shipping continuously.

Kery Cloud connects with GitHub, Vercel, and your preview environments to run browser agents on new changes, then sends screenshots, traces, and repro steps back before merge.

  • Zero-touch QA that runs automatically on product changes.
  • End-to-end GitHub integration.
  • Vercel preview deployment testing.
  • PR-aware testing that understands code changes and intent.
  • Organization and team management with shared projects.
  • More CI and deployment integrations coming.
Early access

Join the Cloud waitlist

Get updates when hosted runs, PR checks, scheduled coverage, and team review are ready.

Why this matters

AI made shipping cheap. Quality is still expensive.

Claude Code, Codex, Cursor, and other agents can create more product changes than a team can manually click through. The risk is not that teams stop building. The risk is that they merge faster than they can verify real user flows.

Reality 01

AI can open more PRs than your team can click through

Coding agents move implementation faster than manual smoke testing can follow. Kery gives every change a browser pass.

Reality 02

The bugs that matter hide in real flows

Login, checkout, forms, permissions, empty states, and mobile layouts are where releases usually break. Kery tests those paths in a real browser.

Reality 03

Fast shipping breaks when QA cannot keep up

The more often you merge, the easier it is for small browser regressions to reach production. Kery catches broken flows before users do.

How Kery works

Give it a product flow. Get back a reproducible bug report.

1

Point Kery at your app

Add a URL, environment, credentials, and the flow you care about.

2

It explores in a real browser

Kery crawls routes, plans steps, logs in, clicks through flows, and observes screenshots plus accessibility state.

3

It returns triaged issues

Visual, functional, and UX bugs come back with severity, screenshots, bounding boxes, logs, and repro context.

0:00
FAQ

What teams usually ask first.

What is Kery?

Kery is an AI browser testing engine. It crawls web apps, runs intent-driven tests in Playwright, and reports visual, functional, and UX bugs with reproducible evidence.

Do I need to write test scripts?

No. You can start with a URL and plain-English intent. Useful flows can later be saved and rerun.

Is this only for simple public sites?

No. Kery supports logged-in products through configured environments and credentials.

Which LLM providers are supported?

Kery supports OpenAI, Anthropic, Google Gemini, and OpenRouter. Bring your own provider keys when running locally.

Does it work behind authentication?

Yes. Kery supports logged-in flows through configured environments and credentials, including form login, Clerk, and Supabase Auth.

Can I run Kery in CI?

Yes. The open-source stack can run in your pipeline, and Kery Cloud is designed for managed PR checks, preview deployment testing, and CI gates.

What can I do with MCP?

Use Kery from MCP-compatible clients like Cursor, Claude Code, and Codex CLI to scan apps, run browser QA, inspect route coverage, and pull bug reports.

Does Kery get smarter over time?

Yes. Kery records useful navigation paths, bug patterns, and known false positives so future runs can focus faster and reduce repeated noise.

Where does my data live?

In the open-source setup, run artifacts live in your own environment. Kery only sends prompts and required context to the LLM provider you configure.

Is Kery open-source?

Yes. The local engine is open-source, built for self-hosting, inspection, and extension. Kery Cloud adds hosted execution and team workflows.

Why have OSS and Cloud?

OSS gives developers the engine and local control. Cloud adds managed execution, team workflows, CI checks, storage, and tenancy.

Community

Join the Kery community.

Ask setup questions, share MCP workflows, compare real browser runs, and follow what lands next.

Join Discord
Get started

Put an AI tester inside your release loop.

Install locally Star on GitHub