Ratifi turns about 260,000 bills from scattered government sources into one plain-language site where anyone can see where a bill stands, who voted how, and what it would cost. For Texas, it goes further than the national data services: the 2027 session appeared on Ratifi the day after pre-filing opened, with hearings, committee votes and fiscal notes read straight from the Legislature's own files. A free public API now lets newsrooms, advocates and school districts build on that data.

The challenge

Legislation is public, but it isn't easy to follow. Congress publishes through Congress.gov, each of the 50 legislatures runs its own website, and votes, bill text and cost estimates live in still other places. Most people meeting a bill for the first time don't know what "engrossed" means, which committee has it, or whether it will ever get a vote.

The goal was a single site that answers those questions in plain language, for every legislature, in both English and Spanish, and stays current through the day without anyone updating it by hand.

What we built

Ratifi was delivered in seven phases, each one live before the next began.

Going Further for Texas

Ahead of the
national sources.

National data services hadn't added the 2027 Texas session when pre-filing opened on November 9, 2026. Ratifi had all 1,360 pre-filed bills the next day, 166 of them on K-12 schools.

🏛️

Hearings and committee votes

Every committee hearing, scheduled and held, with an Upcoming hearings page for the session and the vote count in each committee.

🗣️

Who testified

How many people testified for, against and neutral on each bill, and which organizations took each position. People speaking for themselves are counted, not named.

💵

What it would cost

The fiscal note's five-year cost table, broken down by which state fund pays, read from the Legislature's published files as it asks.

260K
Bills tracked
53
Legislatures: Congress, 50 states, DC, Puerto Rico
2
Languages, 733 translated phrases
79
Automated tests before every release

Problems solved along the way

Real projects run into walls. These are the ones Ratifi hit and how each was handled.

A government website that blocks servers. The U.S. Senate's site refuses requests from server networks, so senators' individual votes couldn't be downloaded. We switched to Voteview, UCLA's research archive of every congressional vote, matched to senators by their official IDs.

A data feed too slow for 50 states. The state data service allows only 500 requests a day on its free tier; loading every state that way would have taken weeks. We built a bulk importer for its downloadable archives instead: 113 files covering every state, DC and Puerto Rico, loaded in under 90 minutes, with the daily feed handling updates only.

Outgrowing the starter database. Past 1 GB of data, the original database couldn't handle several updates at once. We moved to MariaDB with a purpose-built copy tool that kept every record's ID, so no bill link or saved watchlist broke, and checked the result value by value against the original.

A bug only iPhones could see. Tapping "See Texas bills" on the state map did nothing in Safari. We traced it to how Safari moves focus on a tap, reproduced it, and fixed it without breaking keyboard access.

Under the hood

Ratifi runs on one cloud server that we set up and maintain: Python and Django on a MariaDB database, served by nginx over HTTPS. Scheduled jobs pull updates from Congress.gov, Open States, the House Clerk, Voteview, USAspending.gov and the Texas Legislature through the day.

What this means for your project

Ratifi shows how we work: a clear goal, delivered in phases you can see and use, on infrastructure that's documented, backed up and yours. The same skills apply well beyond legislation:

Have data that should
be easier to use?

See Ratifi live, build on its data with the free Ratifi API, or tell us what you're trying to build.