I made a website to convert videos to article (viddo.pro)

Hi HN — we built Factifi, a Chrome extension that does instant fact-checking on anything you’re reading or watching. Please keep in mind, this is just an MVP and will be buggy and fail.

We were constantly pausing podcasts and Substack posts to ask, “Wait, is that actually true?”. After too many 30-minute Google rabbit holes, we decided to automate the work.

What it does

• Extracts claims from articles, blogs, YouTube videos (and even images) • Assigns a verdict + confidence score using live web search, a 200 M-paper research index, and a few proprietary datasets • Flags likely deepfakes and manipulated images • Runs entirely in your browser — only the claim snippets leave your machine for scoring

Status

• Free during early access (no wait-list) • Read more here (factifi.me/blog)

Looking for feedback on

• Accuracy thresholds (where should we draw “unverified” vs. “false”?) • UX friction — what feels clunky? • Ideas for new features

Happy to answer anything in the comments—thanks for taking a look!

Comments (18)

Klonoar · 13h ago

Can you expand more on what "proprietary datasets" means...? Where are the research papers pulled from?

Edit: it's also immediately obvious the number of brand new accounts that are engaging in the comments here with generic comments.

umittal10 · 12h ago

Okay, firstly you we use SEC and a few other datasets from Google and similar datasets. We filter them based on credibility. The research papers are pulled from a corpus called Semantic Scholar which is pretty high quality. We also do ArXiv and PubMed.

ka2429 · 13h ago

Ferrari is losing coz of its strategy. Please fact check it

umittal · 13h ago

Sorry. They’re just ass.

AddySmith2410 · 13h ago

Sounds interesting, can you tell me more about the claim generation?.

umittal · 13h ago

We use some peer reviewed methods to identify questionable claims.

aan85 · 10h ago

How long does it take for a claim to be verified?

shrishti12 · 13h ago

Can it fact check articles, posts etc?

umittal · 13h ago

Yes. Please install. You can see it!

defrost · 11h ago

What does it conclude about the full and complete comment histories of these three accounts?

* https://news.ycombinator.com/threads?id=ka2429

* https://news.ycombinator.com/threads?id=shrishti12

* https://news.ycombinator.com/threads?id=AddySmith2410

Is there are word to describe that behaviour?

umittal10 · 11h ago

Thanks for proving our point for us. People will do whatever they want. Some of us are smart enough to check the truth, but most of us are not. There’s a need for fact checking and hopefully, people can come to a consensus on that.

slater · 12h ago

Can it also check "Show HN" posts with suspicious amounts of simple replies made by suspiciously-new user accounts?

umittal10 · 12h ago

lol we’re trying to prove that the internet is biased.

wredcoll · 11h ago

Biased towards what is the far more important question

umittal10 · 11h ago

Biased towards whatever the hell someone wants to portray. Truth is no longer displayed, what’s displayed is only what you want. HN is one of the better platforms and so you can tell. What about the others where bots dominate?

wredcoll · 10h ago

I'd like it to be biased towards objective facts and not hurting people, how about that?

nonhaver · 11h ago

ah man they got us. we thought it was spam when it was a social experiment all along. now look whos got egg on their face.

ab2345 · 10h ago

Can it be integrated with social media platforms? Since that is where most misinformation is circulated