Gigafact Parser
A civic platform for news professionals to discover what politicians are actually saying.
AI transcribes and analyzes audio and video recordings of public comments by public officials, to surface their claims and talking points.
Used by newsrooms across the U.S.
Cut through to what’s relevant.
It’s painful to listen to a one hour speech or interview just to see if anything useful or important was said. Parser does that work for you. And it is civic-smart and can recognize comments and claims that have journalistic relevance – so you don’t have to wade through needless content.
A prebuilt database of major political figures.
Hundreds of national profile politicians in Congress, Senate and State roles are already present on Parser with thousands of audio and video recordings transcribed and analyzed for claims and major talking points. This database, with contributions by the journalism community, is easily searchable to support research and story development.
A video overview of Parser
How Parser has helped journalists in their political accountability reporting
WISCONSIN WATCH
Fact brief writer Tom Kertscher keeps tabs on the claims that Wisconsin politicians make on AM radio, podcasts, interviews, and more. With Parser, he was able to quickly identify a misleading claim by a congressional candidate accusing her opponent of lobbying for higher gas prices. Tom’s fact brief set the record straight.
“I used to spend hours watching and listening to politicians’ interviews and speeches, searching for claims to fact-check. Gigafact surfaces those claims for me in minutes.”
—Tom Kertscher
TEXAS TRIBUNE
When researching his article on Sen. Ted Cruz’s podcast, Politics Reporter Jasper Scherer used Gigafact’s AI tools to synthesize the key points of dozens of hours of the podcast.
“Gigafact’s tools helped us gain a clear idea of what Ted Cruz says on his podcasts — from the rhetoric he uses about major Democratic figures to the claims he makes about the news of the day — without having to slog through all 500+ episodes he has released. This helped us distill the most important parts while saving us hours we would have otherwise spent listening to old audio.”
—Jasper Scherer
ROCKY MOUNTAIN COMMUNITY RADIO
After Aspen Public Radio hosted a lively hour-long debate between candidates for Colorado’s highly competitive 3rd congressional district, RMCR’s Managing Editor Maeve Conran used our technology to summarize the salient moments in the debate for a website write-up.
“I used Gigafact to process a candidate debate and it was great! … Compared to the Utah gubernatorial debate I had to process last week without Gigafact, it saved me at least 90 minutes of work.”
—Maeve Conran
Want to learn more?
If you’d like to learn more about the platform, or are interested in gaining access, don’t hesitate to get in touch!
Recent updates to Parser
Parser is always being improved by the Gigafact team to ensure it meets the needs of journalists everywhere. Below are some of our most recent updates.
Version 1.7 | Upload from YouTube with a URL & Improved Speaker Identification
- Upload with a URL for most YouTube Videos: When uploading a new YouTube video, you can now paste the URL directly, and Parser will automatically download the MP3 and extract the metadata for you. For non-YouTube sources and very long YouTube videos, the previous upload process is still available.
- Speaker Identification Improvements: Parser’s speaker identification is now 60% more accurate. In tandem, we’ve also enabled all Parser users to override the model’s decision if it gets it wrong, and exposed the model’s reasoning to provide deeper transparency into how it reached its conclusion. The speaker identification popup also now indicates when issues with the transcript may have impacted the identification accuracy.
Version 1.6 | Easier Content Discovery & Favoriting
- Recent Updates: The homepage now features the most recent updates on the platform that might be interesting for you. These are based on your favorited profiles and regions.
- Improved MyFeed: Located in MyParser, MyFeed just got a face lift that now features a paginated view of ALL updates on the platform relevant to your work.
- Improved Favoriting: You can easily revise your favorites with our new process that streamlines finding profiles you’re interested in.
Version 1.5 | Highlighted Claims, View in Transcript, and Improved Navigation
- Consolidated Homepage & Navbar Search: You can now search both subjects and profiles from the same search bar.
- Highlighted Claims: Each source page now features a “Highlight Claims” section showcasing the most important claims identified by our model.
- View in Transcript: On individual profile pages, every claim in each source now has a “View” button—click it to see exactly where that claim appears in the transcript.New Navigation Bar: Every page on Parser now includes a navigation bar, so you can jump to key sections in just one click. Pages also include a search bar so you don’t have to return to the home page to find another profile.