Renderproxy

There was some interest in the product, but not enough to justify keeping it going. A lot of the users were using it as a pure proxy, and hooking up their scraping bots to it to get around IP-blocks, or using it to access blocked sites directly.
There are many good open source solutions out there as well.

The Product Hunt launch was a success in some ways and a failure in others. We got 150 up-votes and were consistently top 7 for the day, and we got some good questions.

Unfortunately, the launch did not result in any paying customers. About 30 people logged in and played with the product as a result of the launch, but none ended up completing the integration.

At this point, you needed to pay to even try the main value proposition of Renderproxy, which in retrospect was clearly a mistake. I changed the free tier to allow you to get all the features (including dynamic rendering), and only divide the pricing tiers by requests / month.

By November I had built something that fulfilled the requirements for Renderproxy. You could go to your dashboard, configure your proxy settings, point your DNS to Renderproxy servers and it would magically implement dynamic rendering for you.

I learned a lot during these few months of development time, including some reason why other solutions hadn't tried the approach I was taking. I decided to continue with launch and let the market decide!

The idea for Renderproxy stayed on my mind for a while, and one weekend I finally decided to build it. The idea was simple: create a service that serves your normal website to users and a pre-rendered version to search engine bots.

Despite some early successes, I realized this was going to be a tough engineering challenge and strapped myself in for the multiple month (side) project it was going to be.

To help improve the SEO of https://0x.org/ we implemented dynamic rendering using Amazon Lambda@Edge and Rendertron. This worked great for us but took over a week of developer time and costs us money in the form of monthly server costs. When looking around, I wish there had been a more general solution.