Commitment to Transparency

Our approach to research and transparency

We believe voters deserve to understand exactly how their ballot information is generated. This page explains our process, our sources, and our limitations — completely.

How we research and build each brief

When you click "Get briefing" on a race, here is exactly what happens.

01
Identify your local races
We use your zip code to identify which metro area you're in and surface the races specific to your area. Your address is never stored or shared.
02
Examine credible sources
Claude (Anthropic's AI) conducts 2-3 targeted searches using real-time web access, pulling from local news outlets, Ballotpedia, official voter guides, and candidate websites.
03
Synthesize multiple sources into a structured brief
The AI organizes findings into a consistent format: office description, race overview, candidate cards with backgrounds and policy positions, endorsements, and a values alignment section.
04
Provide impartial evaluations
The brief is presented without recommendation. The values alignment section connects candidate positions to voter priorities across six issue areas — it never tells you which position is correct or which candidate to choose.

What we research for every candidate

For each candidate, we research their positions across six core issue areas. Not every issue is relevant to every race — a water district board candidate may not have a healthcare position — but these topics guide our research to ensure consistency across briefings.

The EconomyCrime & Public SafetyHealthcareHousing & AffordabilityEducationDemocracy, Voting Rights & Governance

For each candidate, we present their most distinctive policy position in plain language — focused on what they propose to do, not how they describe themselves or their campaign tone.

How we select and present endorsements

Endorsements are presented as informational signals — who is backing a candidate tells voters something meaningful about that candidate's political alignment and priorities. We apply consistent criteria to what qualifies as a relevant endorsement.

We include
  • Elected officials and political figures
  • Political parties and party caucuses
  • Labor unions and worker organizations
  • Civic and advocacy organizations
  • Major newspaper editorial boards
  • Community and neighborhood organizations
We exclude
  • Celebrities and entertainers
  • Athletes and sports figures
  • Private individuals without political office or organizational role
  • Duplicate local chapters when parent organization is listed

We prioritize the most significant endorsements for each specific candidate. A grassroots candidate's most important endorsers may be local organizations — those are listed even if they are smaller than a statewide union. Maximum 5 endorsements are shown per candidate.

Where our information comes from

Ballot Brief uses live web search to pull from publicly available sources at the time each briefing is generated. Primary sources include:

Ballotpedia
Candidate profiles & race data
Local news outlets
Coverage and reporting
Official voter guides
Secretary of State publications
Candidate websites
Policy positions & platforms
Newspaper editorial boards
Endorsements & analysis
Organization websites
Endorsement announcements

How we define metro areas

Ballot Brief uses the U.S. Office of Management and Budget's (OMB) Core Based Statistical Area (CBSA) definitions to determine metro boundaries. CBSAs are the official federal standard for defining metropolitan and micropolitan statistical areas based on population density and economic ties between counties.

When you enter your address, we match it against a database of 4,000+ zip codes mapped to their corresponding CBSA — the same geographic standard used by the U.S. Census Bureau, HUD, and most federal agencies. This ensures our metro definitions are consistent, reproducible, and grounded in official data rather than arbitrary boundaries.

We currently cover 15 major metropolitan areas across the United States, selected based on population size and electoral significance. Our goal is full national coverage by 2027.

Be informed - understand what Ballot Brief can and cannot do

Ballot Brief is intended to be one resource among many to help you make informed voting decisions. We encourage consulting multiple sources, including official voter guides, local news, and candidate websites, before casting your ballot. The following limitations are important to understand before relying on any information presented here.

AI can make mistakes. Ballot Brief uses AI to synthesize information. AI systems can misread sources, miss nuance, or occasionally generate inaccurate information. Always verify important claims with official sources.
Coverage varies by race prominence. High-profile races like Governor or Mayor have extensive press coverage and produce richer briefings. Judicial races, water district boards, and other lower-profile races may have limited web coverage, resulting in thinner briefings or no briefing at all.
Endorsements may be incomplete. We surface the most significant endorsements we can find, but our search may miss some — particularly recent announcements or those from smaller organizations with limited web presence.
Briefings reflect the moment they were generated. Once cached, a briefing does not automatically update when news changes. We review and refresh briefings as election day approaches, but real-time accuracy is not guaranteed.
Local race data is limited. District-specific races (state assembly, city council, school board) vary significantly by zip code within a metro. We are working to add granular district-level data before November 2026.

Ballot Brief is a nonpartisan civic tool built to increase voter participation. We have no political affiliation, are ad-free, and have no interest in which way you vote. If you believe a briefing contains an error, please visit our contact page or email us at hello@ballotbrief.us and we will review it.