GovQL

The problem

I wanted to yell at provide constructive consituent feedback to one of my senators on some of his votes on nominations. Rather than saying, "I disagree with all the times you voted x way," I wanted to be able to say, "Over x nomination votes you've voted y way z times. That's p% - more than all but q senators in your party."

There are some sources for this type of data, but none were kept up-to-date in the aggregate format I wanted. You can download vote data, one XML file at a time, from congress.gov and there's even an API. However, the API didn't cover Senate votes and, because it's a REST API, in order to get all of a certain type of vote for a particular member of congress would involve dozens (if not hundreds or thousands) of API calls.

Around the same time I happened to be playing with GraphQL and it occurred to me that it could be quite useful to build a GraphQL API for government data. It could give journalists and researchers an easier way to dig up interesting insights. The idea languished for a while, but in early 2026 I had time to launch the first version.

What I built

• Give an architectural overview, some screenshots, etc.

Technical decisions

• Keeping it performant on a $6/month DO droplet
• Limiting query complexity
• Postgraphile isn't a panacea - you have to think about the aggregations
• CI/CD - rolling your own