| Core question it answers | Is what's being said actually true? Extracts every verifiable claim and scores it 0-100 against live web sources | Is the face on screen synthetically generated or manipulated? |
| Deepfake / synthetic-face detection | Flags deepfake and AI-generated media signals as one component of a broader analysis | Purpose-built for exactly this. Peer-reviewed, patented PPG method; Intel reports ~96% accuracy in controlled tests and ~91% on real-world 'in the wild' videos |
| Claim-level fact-checking with cited evidence | Yes — per-claim credibility score with source links, plus logical-fallacy flags, on any topic | Not its purpose. FakeCatcher assesses media authenticity, not the truth of statements made in the video |
| Audio and spoken content | Transcribes and analyzes audio; works on podcasts and audio-only sources | Analyzes video pixels only. Independent testing and Intel's own描述 of the method note audio is not part of the analysis |
| Access and pricing | Public web app — paste a YouTube, TikTok, Instagram Reel or podcast link — plus a Chrome/Firefox extension. Free tier; Basic $3.99/mo; Premium $8.99/mo | An Intel Labs research platform, not a consumer download. Runs server-side on Intel Xeon hardware behind a web interface, reached through Intel research and partner channels |
| Real-time throughput at platform scale | Analysis takes time — transcription plus live web cross-referencing on each claim | Results in milliseconds, and Intel has demonstrated up to 72 concurrent detection streams on 3rd Gen Xeon Scalable processors |