Data Sources & Methodology

What CampaignBrief monitors and how.

CampaignBrief uses public and licensed data sources. Base monitoring covers public news, campaign finance filings, official campaign materials, and public platform activity. Premium feeds — paid social APIs, podcast transcripts, local news databases, and broadcast monitoring — are activated only after client approval and configuration.

Base Monitoring

Included in all plans

The following sources are monitored for every active client account. No additional configuration required.

Public news monitoring

News

Monitors public news articles via GDELT and RSS feeds. Covers local news sites, state political publications, and national press.

Data sources GDELT Event Database, RSS feeds from monitored publications
Examples Local daily papers, regional political blogs, statewide political press
Limitation: Monitoring covers public articles only — not private subscriber content. Coverage is broad but not comprehensive for hyper-local outlets not indexed by GDELT.

Campaign finance filings

Finance

Ingests publicly available FEC filings for federal races. Monitors Schedule E independent expenditures, committee disbursements, and PAC registrations.

Data sources FEC public API (federal races), FollowTheMoney for states with open APIs
Coverage Federal races: comprehensive. State races: varies by state.
Limitation: State campaign finance access varies significantly — some states have open APIs, others require manual access or have no structured download available. FEC filings may also lag actual spending by days to weeks depending on reporting deadlines.

Campaign and party websites

Official

Monitors publicly accessible content on official campaign sites, party committee sites, PAC sites, and issue organization sites for message changes, press releases, and published materials.

Includes Campaign news pages, press release sections, PAC issue pages, party committee announcements
Access Publicly accessible pages only
Limitation: Restricted to publicly accessible pages. Password-protected member portals, private donor communications, and internal staff tools are not monitored.

Meta Ad Library

Paid Ads

Monitors political and social-issue ads running on Facebook and Instagram via Meta's Ad Library API — the public transparency archive for ads about social issues, elections, or politics.

Coverage Ads from approximately the past 7 years; active political and issue ads in monitored geographies
Data source Meta Ad Library API (public transparency endpoint)
Limitation: Covers only ads categorized by Meta as being about social issues, elections, or politics. Does not include non-political ads, ads served to audiences built from voter data, or ad performance metrics (impressions and spend are reported as ranges, not exact figures).

Public social platforms

Social

Monitors public posts on accessible social platforms. Bluesky is the most reliably accessible via the AT Protocol open firehose. Reddit via public API. X/Twitter access depends on current API availability.

Platforms Bluesky (AT Protocol), Reddit (public API, rate-limited), X/Twitter (where API permits)
Access type Public posts only, subject to platform API terms
Limitation: Public posts only. Private groups, direct messages, follower-only content, and deleted posts are not accessible and are not monitored. X/Twitter API access is subject to change based on platform policy. Reddit monitoring is subject to commercial API rate limits.

Manual source submissions

Client-submitted

Campaign staff can submit specific URLs, documents, or source tips directly to CampaignBrief. Submitted materials are ingested and analyzed alongside automated feeds.

Best for Local blogs, regional newsletters, opposition research tips, sources not reached by automated feeds
Access type Client-controlled; submitted through secure intake channel
Note: Submitted sources are treated as public documents or documents you have the right to share. CampaignBrief does not accept attorney-client privileged material or documents obtained through unauthorized access.

Premium Feeds

Activated per client

Premium feeds are available as add-ons or in higher-tier plans. Each feed is activated individually after client configuration, onboarding review, and confirmation of source availability for your specific race and region.

Podcast monitoring

Premium

Transcript-based monitoring via licensed podcast APIs. Detects candidate mentions, issue mentions, repeated phrases, and guest appearances across millions of public episodes.

Data sources Podchaser or comparable licensed transcript providers
Best for Catching narratives before they reach mainstream press; tracking regional political podcasts and local talk radio
Note: Coverage depends on podcast indexing by the provider. Not all local or niche shows may be covered. Transcripts may contain errors from automated speech recognition.

Broadcast TV captions

Premium

Monitors television news coverage using the Internet Archive's Television News Archive (GDELT TV Explorer). Covers cable news and local TV news via caption-based text search — not video analysis.

Data sources Internet Archive Television News Archive, GDELT TV Explorer
Best for Detecting when narratives migrate from print and digital into broadcast; local TV pickup monitoring
Limitation: Caption-based only — visual content, on-screen graphics, and B-roll are not analyzed. Caption accuracy varies by broadcaster. Archive availability may lag broadcasts by hours.

Local news databases

Premium

Subscription-based local news monitoring covering hyperlocal and regional publications not indexed by GDELT or accessible via public RSS. Extends base news coverage for races where local media is the primary arena.

Best for County-level races, state legislative districts, and media markets with strong community newspaper presence
Availability Coverage varies by region; confirmed during onboarding
Note: Regional coverage varies. We confirm specific publication availability for your race during onboarding before activating this feed.

State campaign finance

Premium

State-level campaign finance monitoring for states with accessible APIs or downloadable data. Tracks state PACs, candidate committees, and independent expenditures not covered by FEC.

Access model Automated for states with open data (e.g., FollowTheMoney partners); manual review for states with limited access
Availability Varies significantly by state; confirmed during onboarding
Note: State-level access varies significantly. For states with limited or no structured data export, monitoring may be manual or unavailable. We confirm state-level availability during onboarding for every client.

Influencer and podcaster tracking

Premium

Custom monitoring of specific named hosts, reporters, local columnists, or public social accounts relevant to your race. Requires a watchlist configured during setup.

Best for Tracking key validators, local opinion leaders, and accounts that regularly drive district-level narrative
Configuration Watchlist submitted and reviewed during onboarding; public actors only
Note: Limited to public actors acting in a public role. Personal accounts, private individuals, and non-public content are not eligible for watchlist inclusion.

What CampaignBrief does not monitor.

The following categories are explicitly outside the scope of CampaignBrief monitoring. This is not a policy gap — it is a deliberate design choice. CampaignBrief is a public-source intelligence tool, not a surveillance platform.

How phrase convergence works.

Phrase convergence is detected when the same or similar wording appears across multiple independent public sources. CampaignBrief flags convergence when the same phrase, claim, or message frame appears in two or more unaffiliated source categories within a defined time window.

The system tracks whether a phrase originated in a single source type (e.g., only social media) or has migrated across source types (e.g., social media to news to paid advertising). Cross-source-type convergence is weighted more heavily than same-source volume, because the same phrase appearing in a PAC ad, a local news article, and a podcast episode is a stronger signal than the same phrase appearing in dozens of social posts from a single account cluster.

What it means The same language is spreading publicly across independent sources. The phrase, claim, or frame has crossed source-type boundaries and may be gaining traction.
What it does not mean Legal coordination, foreign involvement, or campaign finance violation. CampaignBrief does not assert legal conclusions. Human review is required before characterizing phrase convergence as evidence of coordinated activity.

AI analysis.

CampaignBrief uses AI (OpenAI where enabled) to cluster narratives, identify repeated phrases, assess source migration, and generate daily brief summaries. AI analysis is applied to public source text excerpts only — not to personal account data, voter information, or client communications.

AI outputs are labeled as AI-generated throughout the platform and require human review before campaign action. AI analysis can miss context, misread tone, produce errors, or reflect limitations of the underlying model. It is a drafting and awareness tool — not a final authority on what a signal means or what action a campaign should take.

AI is used for Narrative clustering, phrase identification, source migration assessment, daily brief drafting, and signal summarization from public text.
AI is not used for Generating campaign content, producing synthetic media, making legal conclusions, targeting voters, or making decisions on behalf of the client.

Source confidence levels.

Every signal card in CampaignBrief displays a confidence level. This reflects the quality and corroboration of the underlying sources — not how important the signal is or how much attention it warrants.

High confidence

Multiple independent credible sources with a verifiable public record. Examples: an FEC filing, an official campaign statement published on a candidate website, a published news article with a named byline in a credentialed publication, or a confirmed Meta Ad Library entry with a named buyer. High-confidence signals have enough sourcing that a campaign should treat them as confirmed public information requiring no additional verification before discussion or response preparation.

Medium confidence

1–2 sources with a credible source type, but limited independent corroboration. Examples: a single local news article, a podcast mention without a written source to cross-reference, or a social post from a credible organizational account without other confirmation. Medium-confidence signals are directionally reliable but warrant verification before any operational response.

Low confidence

Single source, anonymous source, social post, or source type with limited accountability. Examples: an anonymous blog post, a single social account post without supporting sources, or a community listserv item. Low-confidence signals are included in CampaignBrief so staff can monitor early-stage narratives. They should be treated as alerts for further investigation — not confirmed facts. Acting on or amplifying a low-confidence signal without independent verification is a campaign risk.

Want to see CampaignBrief monitoring your race?

Our team reviews every access request and confirms source availability for your specific race and region before onboarding. We do not run a self-serve signup — every client account is set up with a configuration review.

Request Access

CampaignBrief provides public-source monitoring, evidence organization, and analytic summaries for campaign operations awareness. All monitored content is drawn from publicly available sources. CampaignBrief does not create campaign communications, target voters, generate synthetic media, assert legal conclusions, or act as a substitute for qualified legal, compliance, or campaign counsel. Human review is required before any campaign action. Source availability, premium feed activation, and state-level coverage are confirmed individually during onboarding. Operated by SapienCX LLC, a Virginia limited liability company. See our Terms of Service and Privacy Policy for full details.