🚦🕵🏻♂️ Fact-Checker Button for Transcripts
L
Lucho
1. Motivation
During meetings, participants frequently share data, statistics, market claims, history and technical information that may require verification. Currently, users must either web search during meetings or manually fact-check statements after meetings, creating delays, distractions and potential mistakes. A built-in fact-checking feature would enable immediate verification of transcripts, enhancing decision-making quality and meeting credibility.
2. Feature Request
Implement a
Fact-Checker Button
that generates lightweight verification reports for recorded meetings. When activated, the system would:- Extract verifiable claims from transcripts (dates, statistics, historical facts, scientific statements)
- Cross-reference information against reliable public databases
- Produce a brief report highlighting supported, refuted, or conflicting claims
- Display confidence levels with source citations
The report should be concise and actionable, not exhaustive like academic fact-checking tools.
See attached example below
3. Implementation
Integrate a backend evidence-based verification system using free public data sources, for instance:
- CIA World Factbook(github.com/factbook) for geographic and demographic data
- Semantic Scholar API(github.com/JackKuo666/semanticscholar-MCP-Server) for scientific claims
- ClaimAI framework(github.com/BharathxD/ClaimeAI) for claim extraction and verification architecture
The system would process transcripts post-recording, identify factual statements using NLP, query relevant databases, and generate a summary report showing verification status (Supported/Refuted/Conflicting) with percentage confidence and source links.
This feature would differentiate Hedy as a professional tool for high-stakes meetings where accuracy and ethics matters, avoiding misinformation.