Trudl

We decided to launch a forever free plan through a desktop application and add premium features later on. We built the desktop app using Electron JS, which seemed straight-forward but was anything but - more on that in a blog later.

We were able to create a test-running algorithm that worked on most test flows. Additionally, we built a CSS-selector feature that would make the test flows more robust (https://docs.trudl.io/css-selectors/adding-a-css-selector).

We realised that we had to create a test-running algorithm that would work on most of the websites and all types of buttons, links, drop-downs, form elements etc.

We also had to build a lot of features - editing tests, asserting elements, adding wait time etc - that not only were requested by our early users but also were in our product roadmap.

So we decided to spend the next few months to work on the algorithm and build some necessary product features.

We reached about 40 private beta users. Only about 30-40% of the users who had signed up for early access responded - one mistake we made was to not keep in touch with them and send them regular updates.

Also, since for each user we were taking too long to test and make Trudl work on all their website flows, we decided that it was not scalable and went back to the drawing board.

We built out our MVP in about 4 months and decided to launch private beta of Trudl to get user feedback.

We had about 70+ people who had signed up for early access - we decided to personally mail all of them.

We also decided to use LinkedIn and Twitter to get more private beta users.

We brainstormed for a couple of days, looked at best landing page examples and created a landing page - https://trudl.io - to get beta signups. One feature we really like in our landing page are the GIFs that explain our product very well.

If you have any suggestions to improve our landing page - let us know by replying to this thread https://www.indiehackers.com/post/fdf0ecc2f2

We wanted to choose an area of automation that was technically simpler than building self-driving cars. And a go-to-market easier than convincing companies to use conversation-bots for sales. And one where we wouldn't face regulatory challenges as we would have had we decided to fly drones.

Regression testing, as I had experienced, was a major pain point in my previous companies and automating tests and maintaining them was difficult.

We were convinced that a no-code automation product that would solve this problem would add a lot of value to tech companies.

Praveen and I had first met through a mutual friend in 2015 and had been acquainted ever since. As both of us wanted to build a SaaS product, we caught up and discussed a bunch of ideas.

The prominent theme of our discussion was automation that would create massive value in the future - drone delivery, self-driving cars, bots that can converse etc.