Test apps built with Devin before you merge the PR
Devin ships a full pull request where the build passes but a real journey can break. Kane CLI runs that branch in a real browser, verifies each step in plain English, and returns a pass or fail. Devin can run it. Free to install.
or read the documentation
Why test apps built with Devin
Devin, from Cognition, is an autonomous software engineer: hand it a ticket and it scopes the work, writes code, runs the build, and opens a standard pull request. You review the diffs, often several parallel Devins at once.
Large diffs break review. The build passes, but agent faults are semantic, not syntactic: a missing API method, a behavior-changing refactor. Kane CLI runs the branch in a real Chrome browser and names the step that breaks.
Kane CLI is agent-native, so Devin can verify its own work. Point it at testmuai.com/kane-cli/agents.md and it runs Kane CLI with the --agent flag: ship the ticket, run the journey, read the result, fix the failure, open the PR.
What Kane CLI tests in your Devin app
The feature on the ticket, the journeys the diff touched, and the runtime bugs review misses.
The ticket Devin shipped
Walk the exact journey the ticket described, from entry point to new behavior, and confirm the feature Devin opened a PR for works in the browser, not just that the diff looks right.
Regressions the diff hides
Devin's PRs are often large diffs that touch shared code. Re-run the journeys around the change so a refactor or renamed function does not quietly break a flow nobody re-tested.
Semantic bugs at runtime
Lint and build pass while a fabricated API call, a wrong config value, or a subtle logic change only fails when the app runs. Kane CLI exercises the live app and catches what static checks wave through.
Auth and protected routes
Sign up, log in, and confirm sessions, protected pages, and role-based access still behave after Devin's change, the edge cases a quick PR review rarely clicks through.
Forms, data, and core actions
Submit forms, run search and filters, complete the core workflow, and verify the data lands where it should, so a passing build never masks a broken submit.
Console errors and broken assets
Crawl the running branch for console errors, broken links, and missing assets that only surface in a real browser, before they reach the reviewer or production.
Review Devin's PR with proof, not a hunch
Start in your terminal
Validate on the cloud
Release with confidence
Verify what Devin ships before it merges
Kane CLI and KaneAI share the same automation engine and dashboard.
Verify the PR before you merge it
Devin's output is a standard pull request. Kane CLI verifies the running branch in a real browser so you review a proven feature, not a plausible-looking diff that breaks on first click.
Every parallel session checks its own work
Run several Devins at once, or delegate to managed Devins, and each drives Kane CLI on its own branch via the --agent flag, reads the result, and fixes the break before its PR lands on your desk.
Evidence attached to the diff
Every run produces a persistent video, a step trace, and a replay link you can drop straight into the PR, so proof that the flow works travels with the change Devin shipped.
Test your Devin branch in three steps
Install Kane CLI
Run npm install -g @testmuai/kane-cli and sign in with your TestMu AI account. Nothing to wire into the repo Devin is working in.
Run the branch, point Devin or yourself
Start the app from Devin's PR branch, then send Devin to the guide to drive Kane CLI with the --agent flag, or run flows yourself against the running URL.
Describe the flow and verify
Write the journey the ticket asked for in plain English. Kane CLI drives a real browser, verifies each step, and returns a pass or fail with evidence.
Get Started With Kane CLI
🎉 Launch offer: Bonus credits for the first 3 months on paid plans
Choose the right plan for you
Local test authoring via CLI
Auto-heal & vision
View test cases on UI
Test Manager
Free
$0
/month
200 Credits
Resets in every
30 days
Starter
$19
/month
2000 Credits
Launch: 4,000 Credits (+100%)
Bonus for first 3 months
Pro
$99
/month
10,000 Credits
Launch: 15,000 Credits (+50%)
Bonus for first 3 months
Enterprise
Get access to solutions built on Enterprise-Grade Security, Privacy, and Compliances.
Need more credits?
Got a bigger use case in mind?
Let’s talk
Choose the right plan for you
Free
$0
/month
200 Credits
Resets in every
30 days
Starter
$19
/month
2000 Credits
Launch: 4,000 Credits (+100%)
Bonus for first 3 months
Pro
$99
/month
10,000 Credits
Launch: 15,000 Credits (+50%)
Bonus for first 3 months
Enterprise
Get access to solutions built on Enterprise-Grade Security, Privacy, and Compliances.
Need more credits?
Got a bigger use case in mind?
Let’s talk
Get the technical rundown
Documentation
Everything you need to install, configure, and run Kane CLI in under 2 minutes.
Frequently asked questions
Spin up the branch from Devin's pull request, run the app, and describe the journey in plain English, for example "open the ticketed feature, complete the new flow end to end, and confirm the existing dashboard still loads." Kane CLI drives that running app in a real Chrome browser, checks each step, and returns a clean pass or fail with video evidence. Run it yourself before you review the PR, or let Devin run it in the same session where it scoped the ticket.
Devin is an autonomous software engineer that picks up a ticket and ships a full pull request, often a large diff that spans the stack. Lint and the build pass, the diff reads plausibly, and that is exactly the failure mode that slips through. Research on agent-written code shows most faults are semantic, not syntactic: a fabricated API call, a config option that does not exist, a refactor that quietly changes behavior, a flow that only breaks at runtime in the browser. Human review degrades on large diffs, so a browser-level check on the running app catches what code review misses.
Yes. Kane CLI is built for AI coding agents. Point Devin at the guide at testmuai.com/kane-cli/agents.md and it installs Kane CLI, runs flows with the --agent flag, reads the structured results, and fixes what failed before opening the PR. With parallel Devin sessions or managed Devins, each one verifies its own running branch the same way.
The new flow the ticket asked for, plus the journeys around it the diff might have touched: sign-up and login, protected routes, forms and validation, search and filters, checkout, and navigation. It also surfaces console errors, broken links, and missing assets that only appear when the app runs. It checks each step, not just the final screen, and flags the exact point where the running app breaks so Devin fixes the right thing.
Yes. Devin already runs CI checks with GitHub Actions before merging, and Kane CLI fits the same gate. Authenticate with your TestMu AI credentials, pass --headless and --timeout, and gate the pipeline on the exit code: 0 on pass, 1 on fail, 2 on setup or auth errors, 3 on timeout. Run your key flows on every Devin PR so a regression in a large diff never reaches main.
The CLI is free to install and use. Local runs are free; cloud runs on the TestMu AI grid are billed against your TestMu AI plan. Start on the free tier and verify the branch Devin shipped end to end without a credit card.
Teach your agent to test what it builds
Point Devin at the Kane CLI guide and it installs, authenticates, and verifies the running app it built in a real browser on its own, before you review the change.
Point your agent to: testmuai.com/kane-cli/agents.md