The summer that keeps on giving:

• You now interact with Octomind via their CLI now. You can create, execute, and run tests, debug test cases locally, launch the private location worker and generate tests in batches.

• You can use teardown to isolate your tests - we added teardown functionality to the steps editor. Teardown allows you to configure a test that cleans up artifacts created during test execution automatically. You can set it up using the app s interface or programmatically via a javascript step.

• More news in the August product update https://octomind.dev/blog/produc...

You never signed up to be a prompt engineer - and now you don t have to. We ve tuned the Agent to clean up and enrich whatever natural-language prompt you give it. The Agent expanded that prompt into a structured prompt with three clear sections:

1. Test INTENT a description of the user journey

2. INSTRUCTIONS the concrete interactions (steps) to perform

3. EXPECTED OUTCOME the final state, ready to become an assertion

And if the autotuned prompt drifts, you can jump in, edit it, and steer the Agent back on track.

Converse with the Agent to direct Octomind! This goes far beyond a glorified AI docs search. Via MCP, we enabled our agent to launch generation workflows, create items, change settings, query and execute tasks directly from the chat window. All the things you might need too much time to figure out and do manually in Octomind app.

• Agent, create these tests and run them

• Agent! set up a new environment for me

• How do I set up authentication for my login test?

Octomind was, is and will always be AI-first. AI is THE technology to enable a substantial increase in testing productivity. What we learned in the process of building it is how people interact with AI - where its assistance is the most useful and where you need to plug in a human to propel the Agent in the right direction. This informed the philosophy behind our new test recorder feature.

We have built a Recorder that enables our users to co-create and co-edit in collaboration with the Agent in all test creation and adjustment workflows. It gives you full control over the test generation process and speeds things up where AI makes little sense.

you might have noticed over the past few weeks, that out test reports look differently. Failed tests are clustered, test run details is packed with more information and new buttons appeared leading to new features. Let me explain.

From the start we knew that after test generation and execution, we would have to tackle test maintenance next. It's the reason why many teams decide to stay away from end-to-end testing entirely. It's time consuming and w

annoying. So we made it our goal to cut time spent on fixing tests. By a lot.

This Wednesday, we'll be rolling out an update to the submission flow. Historically, at the end of the submission flow, you could either Launch Now or Schedule for a future time. Going forward though, we're going to replace the Launch Now option with a new Create Draft option.

This change is based on consistent feedback we've gotten from Makers who either accidentally launched their product too early or who wanted to be able to see a complete version of their launch and collaborate with their team on it before scheduling.

Creating a Draft

Creating a draft will give you the opportunity to see a complete version of your launch without needing to select a launch date. It also means that you'll be able to share and collaborate on the launch with your fellow Makers, something that's not possible when the launch is in the submission flow stage. These new draft launches won't be indexed on the site, so you don't have to worry about others finding it. Whenever you're ready, you can schedule the launch for the date of your choice.

Drop a link to your product launch in the comments.
Let's support dev-first products
Upcoming launches:

• Octomind

• Prisma

• Prelude