On Amazon, a product that comes in several colors or configurations can carry a different price for each one. If the variant selector updates the photo but keeps the old price, the wrong item quietly ends up in the cart. What Kane CLI checked on ShopKart: > picked Arctic White and confirmed the Wireless option was priced at $84.99 > switched to Wired and saw the price drop to $69.99, with the selection label updating to match > added the headset to the cart and matched the variant and price in the confirmation Kane CLI reads the price and the selection label together at every step, so if the page updated one and not the other, the run would fail. Verdict: PASSED, 3 of 3 steps in 131s, full run in the video. Free to start 🔗 testmuai.com/kane-cli/?utm_s… #TestAutomation #KaneCLI

kane-cli 0.8.17 🏗️ 📡 Non-TTY output is versioned NDJSON. Parse, don't scrape. 🧵 Better progress reporting for test runs. 🔗 Job URL up front on --remote. No waiting to watch. 🪟 Windows skill installs fixed. Tarballs extract again. Streams you can parse github.com/LambdaTest/kane-c…

If you run or build a product that takes payments, verify the dashboard lookup on every release. A charge can succeed at checkout and still show the wrong amount or status on the dashboard, and support is usually the first to hear about it. What Kane CLI checked on Stripey: > completed a $240.00 checkout and read the Payment ID from the receipt > searched the Payments dashboard for that ID and found a single succeeded row > opened the payment detail and matched the amount, status, and Payment ID Since the ID from checkout becomes the search term, every later check is about that same transaction. Verdict: PASSED, 4 of 4 steps in 156s, full run in the video. Next in our series on flows a business can't afford to break. Free to start 🔗 testmuai.com/kane-cli/?utm_s… #TestAutomation #KaneCLI

kane-cli 0.8.16 🎯 🔁 Replays use this run's values, not author-time. 🧱 Contracts match endpoints, not one run's IDs. 🪟 Window size = real viewport, attached browsers. ⌨️ Key combos use Cmd on macOS, Ctrl elsewhere. Replays that tell the truth github.com/LambdaTest/kane-c…

Terminal-first is CI-first. If a test can't run from a shell script, it can't run in CI. 🖥️ 🔁 #KaneCLI

$ 𝖪𝖺𝗇𝖾 𝖢𝖫𝖨 --𝗍𝗂𝗉 𝟤𝟥 💡 Name the run and it becomes a test file. kane-cli run "go to /plans, store the Fiber 100 price as 'fiber100_price', assert it is '₹799/mo'" --name fiber100-price One flag. Kane writes .testmuai/tests/fiber100-price_test.md plus an output folder with the result and the recorded steps. No copy-paste into a file later. #KaneCLI

If you run or test an ecommerce store, search is the first flow to verify. It can break without an error or an alert, and the only sign is shoppers leaving. So we gave Kane CLI a single objective written in natural language, with no selectors and no test script, and it ran the flow in a real Chrome browser on ShopKart. Here is what it checked: > searched for "headset" and confirmed [WHAT MADE THE RESULT RIGHT] > opened the product page and checked the title, brand, and price > confirmed the buy box showed In Stock, Add to Cart, and Buy Now During the replay, the recorded search click no longer matched the page because, so Kane CLI worked the step out again from the objective and carried on. Matching is confidence-scored, so an ambiguous element gets flagged instead of clicked, and a real bug isn't healed over. The run passed all 3 steps, and the full recording is in the video. This is the first in our series on flows a business can't afford to break, and is next. Kane CLI is free to start 🔗 testmuai.com/kane-cli/?utm_s… #TestAutomation #Ecommerce #KaneCLI

kane-cli 0.8.15 ⚔️ ✨ Data-driven runs 🔍 Uploads that find the right file 🌟 Variables that warn instead of blocking 🔁 Replay that heals itself Replay, retry, recover. github.com/LambdaTest/kane-c…

kane-cli 0.8.14 🚀 📤 Share a context store with your team ✨ New store housekeeping 🌟 Attach to existing appium sessions 🧪 Tests can branch and loop context, shared. github.com/LambdaTest/kane-c…

--remote now works for mobile on Kane CLI. Add one flag to a mobile run, and it goes straight to HyperExecute. > Runs on Android emulators and iOS simulators hosted on HyperExecute > Nothing to install or keep running on your own machine > Only devices that actually boot show up in the list Web already had it. Mobile has it now. Kane CLI 0.8.13 is live 🔗 testmuai.com/kane-cli/?utm_s… #KaneCLI #HyperExecute

kane-cli 0.8.13 📱 📤 Anyone can now run emulator / simulator ☁️ --remote runs emulator and sim tests on HyperExecute. 🔗 App id forwarded automatically. Runs grab it. 🍎 Sim / Emu catalog cut to what mac images boot. Local build, remote devices. github.com/LambdaTest/kane-c…

kane-cli 0.8.12 🔒 🚫 Unresolved variable? The run refuses before dispatch. 🧩 Design sessions declare variables and write the stubs. 🔗 A stored value carries into the next section by name. 📸 Device runs upload screenshots again. Fill it, or it won't start. github.com/LambdaTest/kane-c…

The Kane CLI Hackathon is done, and the results are in. 🚀 🥇 Sam Felix (@SamFelix0512 ), Founding Engineer (Full Stack) at Nitrane Systems 🥈 Abhinav Pangaria (@abhinavpangaria ), Founder at PayZoll 🥉 Christine Lall, QA and Test Manager at Cayman Islands Government A founder, a founding engineer, and a QA lead in government, all landing in the same place from very different starting points. Congratulations to the three of you, and thank you to everyone who submitted and made the judging difficult. 🙌 Want to build what they built? Start here: testmuai.com/kane-cli/?utm_s… #KaneCLI #TestMuAI

kane-cli 0.8.11 🚀 🔁 --author re-authors a test in place, identity kept 🧩 *_module.md imports a cloud module 📌 DB/API/JS steps stay on their tape through heal 🔢 {{var.path}} and {{rows[0].id}} resolve end to end Same tests, fewer surprises. github.com/LambdaTest/kane-c…

Normally, this is where the maintenance ticket gets written. 📝 A minor frontend code change breaks a test step ➔ test crashes ➔ you spend 30 minutes updating brittle selectors. We built Kane CLI so you never have to do that again. Check out this 30s demo where Kane detects the breakage, triggers its retry engine, re-authors the plan on the fly, and passes the test without human intervention. Try the new age auto-healing 🔗 testmuai.com/kane-cli/?utm_s… #KaneCLI

$ 𝖪𝖺𝗇𝖾 𝖢𝖫𝖨 --𝗍𝗂𝗉 𝟤𝟤 💡 Always set a timeout. 𝗄𝖺𝗇𝖾-𝖼𝗅𝗂 𝗋𝗎𝗇 "𝖺𝗌𝗌𝖾𝗋𝗍 𝖼𝗁𝖾𝖼𝗄𝗈𝗎𝗍 𝗐𝗈𝗋𝗄𝗌" --𝗍𝗂𝗆𝖾𝗈𝗎𝗍 𝟤𝟢𝟢 If a page hangs, the agent would sit there forever. A timeout makes it fail fast so you can move on. #KaneCLI

kane-cli 0.8.10 ⚔️ 📤 Mobile code export is here (Python only) ✨ More accurate mobile exports 🐛 Scroll and until-loop fixes 🔁 Mobile replay reliability Replay, retry, recover. github.com/LambdaTest/kane-c…

Fable 5.1 makes verification MORE urgent, not less. Multi-day autonomous coding sessions mean multi-day gaps between "written" and "clicked through" 🤖 Give your agent a real browser 🌐 #KaneCLI

API testing with no Postman collection to maintain. One sentence, one negative path: POST to an insurance portal's login API with an email that was never registered, confirm the response is 401 and the error field reads "invalid credentials". Real Chrome. Request sent, response read, verdict returned. The run passed because the API refused correctly. Point it at your own endpoints, free to start 🔗 testmuai.com/kane-cli/?utm_s… #KaneCLI

$ 𝖪𝖺𝗇𝖾 𝖢𝖫𝖨 --𝗍𝗂𝗉 𝟤𝟣💡 Put settings at the top of the test file. # header: mode, max-steps, timeout List the mode, step limit, and timeout in the file header. You stop repeating those flags on the command line every run. #KaneCLI