Data Sources
Every record traces back to an official source.
Republiq does not scrape, repackage, or re-host third-party news coverage as the basis for civic claims. Analyses are grounded in authoritative APIs operated by the same institutions that originally published the records — Congress, the FEC, the federal courts, and a small set of well-known civic data nonprofits.
Live ingestion
Sources currently powering production features.
Congress.gov API v3
Library of Congress
api.congress.gov/v3
What we use it for
Federal bills, sponsors, amendments, committee actions, roll-call votes.
Where it shows up
Bill enrichment, legislative tracking, prediction-market resolution.
CourtListener / RECAP API v4
Free Law Project
courtlistener.com/api/rest/v4/search
What we use it for
Federal court opinions, dockets, and PACER-mirrored filings.
Where it shows up
Research Suite, legal-context citations.
OpenFEC API v1
Federal Election Commission
api.open.fec.gov/v1
What we use it for
Campaign finance: candidate fundraising, committee filings, expenditures.
Where it shows up
Politician profiles, financial-disclosure context.
Vote Smart API
Vote Smart (nonprofit)
api.votesmart.org
What we use it for
Candidate positions and voting records across 14 spectrum topics.
Where it shows up
Stance synchronization on politician profiles.
NewsAPI v2
NewsAPI
newsapi.org/v2/top-headlines
What we use it for
Real-time news headlines for framing and hype-index analysis.
Where it shows up
News Lens framing analysis.
X API v2
X Corp
api.twitter.com/2/trends/by/woeid/23424977
What we use it for
US trending topics and politician public statements.
Where it shows up
Politician mindset snapshots, trending civic topics.
Wikipedia API
Wikimedia Foundation
en.wikipedia.org/w/api.php
What we use it for
Plain-text enrichment of ballot measures (cached 30 days).
Where it shows up
Ballot measure context with fallback to official Secretary of State text.
On the roadmap
Sources scoped, in some cases connector-stage, but not yet production-active. Listed here so the product surface stays honest.
OpenStates
State-legislature bills and voting records. Connector exists; not yet wired to primary product flows.
Voter Information Pamphlets
State-by-state pamphlet ingestion. Currently admin-uploaded; automated ingestion planned.
Source-handling policies
- ◆Records are pulled from official APIs at ingestion time and stored with the upstream identifier so any analysis can be traced back to the canonical document.
- ◆Bill text passed into AI analysis is truncated to a fixed length to keep downstream behavior predictable; the citation always points to the full upstream document, not the truncated excerpt.
- ◆External enrichment data (Wikipedia, news headlines) is cached with a defined TTL to limit drift between what the model saw and what users see.
- ◆Republiq does not aggregate user-generated social content as a source of civic claims. X is used only for trending topics and accounts of named public figures.
See the records for yourself.
The beta exposes the same APIs through a single, neutral surface.