Gitgrep
Regex search all your git repos in milliseconds.
I'm sure I'm not the only one here obsessing over their Product Hunt launch. My biggest concerns are:
- Does the landing page make sense?
- Does the landing page convert?
- What do I use for the Product Hunt launch thumbnail?
I'm not kidding about (3). I have little to no graphic design experience and want an elegant animated GIF that conveys what the product does. As soon as I have this, I'm ready to launch I tell myself. Or maybe that's just my latest excuse for not pulling the trigger. Stay tuned, we should know very soon.
On Sunday morning I was just browsing Hacker News and noticed that the top story was "How Google Code Search Worked: Regex Matching with a Trigram Index (2012) (swtch.com)". Now this was very fortunate as Gitgrep (based on Etsy/Houndd) is also a Regex code search using a trigram index. Talk about co-incidence.
I had just worked all day Saturday and thought I would have some relax time, but instead I posted a comment on that story with a link to my as-yet-unfinished landing page, gitgrep.com. The very next thing I had to do was add Google Analytics to the landing page. As I saw visitors from all around the world, I spent the rest of the day working as hard as I could to get as much of the landing page finished and working as possible. I managed to get the checkout, account provisioning and welcome email all connected.
As I watched the traffic fall off, none of the 100+ visitors signed up. Of the visitors that didn't immediately bounce, the time on page was around a minute showing that they were reading the content. Also, I didn't add GA to the demo page so they may have gone there but registered as a bounce.
Main takeaway is that I don't know what they understood, liked, or didn't get/like on the landing page. Figuring this out is the big ticket.
Having decided that each account would have its own machine instance meant that as soon as someone signs up, a new machine would have to be prepared and customized for that user. The work to automate this was more involved than anticipated.
I started by manually making a demo site using some Vultr credits that I had for side-projects. While working, I realized that there was a piece missing for instances to get even the simplest metadata about itself, so I got briefly distracted into making a Vultr metadata service, similar to what I'm used to having available on EC2.
Around this same time, I was juggling domain names for projects and decided to have a common one as an umbrella one for administration and miscellaneous projects thus QuickX.app was born.
I started small by adding https (TLS) support to my fork of etsy/hound as well as recognizing or rejecting a JWT authentication http header. Work continued from there to add the login/logout as well as navigation UI. Integration with the SaaS account administration was the next thing.
Releases v0.2.0s (TLS) and v0.2.0t (top navigation+login UI)
Just around this time, I was wrapping up launching my first product Quicklog.io. It is a more niche product that will have a longer marketing/discovery period, so I decided to start on something else right away.
It then struck me that the very thing I've enjoyed using so much would be useful to any developer. I started the very same day to get to work on taking the MIT licensed etsy/hound project into a SaaS offering. I thought about what might be important and security was at the top of my list when using it myself so I used that to guide the technical architecture of the product.
I first discovered the etsy/houndd project on Github back in 2016. I set it up at the startup where I was working and it was big hit with the devs. It only worked in the office as I had deployed it on a host on the office network for security. The project as it was didn't offer login or security features of any kind. At the time I was just happy to be able to search all the company's code repos and didn't think any more of it.
Comment