I hope you’re having a good week so far. No big intro today, straight to the new stuff.
Custom variables. You can create your own custom variables now. Reference them in prompts using '$' (e.g. $firstname) or in the locator picker by adjusting the ‘enter text’ step. Do you have an existing variable list? No problem. You can upload it via a json file. 💲
Environments. We support multiple environments with separate test accounts, URL and OTP settings now. It’s a structured way to isolate configurations for different deployments, customize testing settings specific to each deployment (e.g. user credentials) and perform tests without interfering with production data or workflows. More. 🌿
Even Debugtopus (our local runner) is now environment aware ;)
Native support for OTP. We support logins that require a 2nd factor using a one-time-password (OTP). It’s configurable in settings and works for local execution too! 🧑💻
Pipeline blocking. If you develop using CI pipeline, you can have Octomind check if you broke a new build and block it. When integrated, it can block the pipeline by polling the resulting test report until the status is 'PASSED' with a call to the test report API. 🛑
Longer retention of snapshots. We will retain your snapshots for 6 instead of 2 weeks. You’ll see the snapshots of the last successful run much longer now. You needed more time and we listened. 📸
Correct tracking of animated elements. To edit a test case, it's really important that our locator picker tracks the element position correctly at all times. Even when animated. It’s key to selecting the right element to interact with. 👇
We have a new external API for retrieving test reports. If you use Octomind to monitor your app via external application you’ll be able to see previous runs and compare. ↔️
We introduced custom variables twice. At first, we introduced concepts difficult for users to understand. So we went back to the drawing board and implemented a more intuitive version.
We are teaching the agent to notice if it got stuck in a loop and abort. Loop breaking is surprisingly complicated. It’s not easy to tell a doom loop from a productive loop apart. You don’t want to stop the agent generating too early.
Shared authentication state. Login is a required dependency for most SaaS app tests. In the past every test logged in separately overloading authentication channels, triggering bot alarms at auth providers, and creating inefficiencies for parallel testing.
When you run your Octomind test suite now, we log in just once. All the tests using the login as dependency share the auth state.
---
I’ve just realized I recall the customers who asked for 8 out of 10 of these features. Good to know we develop Octomind in lockstep with our users.
That said, keep suggestions and feedback coming, please.