Launching today

Facts...No Bullsh*t

Launching today

Stop BS in real-time with AI that fact-checks as you listen

68 followers

Stop BS in real-time with AI that fact-checks as you listen

68 followers

Visit website

Realtime Voice AI

Most fact-checkers make you wait — copy text, paste links, hope for results. By then, the damage is done. We built fact-checking that happens while you listen. Turn it on during a podcast, meeting, or call. Our AI detects verifiable claims in real-time and instantly checks them against reliable sources. What's different: Truly real-time — verified as words are spoken Works with any audio source Clear TRUE/FALSE verdicts with sources Know what's true the moment it's said.

Free Options

Launch tags:Artificial Intelligence

Launch Team / Built With

Migma — Lovable for Email

Lovable for Email

Promoted

Facts...No Bullsh*t

Maker

📌

Hey Product Hunt! 👋 I built this out of pure frustration. I was listening to a podcast where the host confidently stated something that sounded off. By the time I paused, Googled it, and found it was completely false — I'd already missed 10 minutes of content. That moment stuck with me: why do we still fact-check manually in 2026? So I built a tool that listens with you and verifies claims as they happen. No copy-pasting, no context switching. Just real-time truth. It's been a game-changer for me during meetings, podcasts, and even political debates. I'd love to hear: What audio sources would you use this for? Any features you'd want to see added? Happy to answer any questions. Thanks for checking it out! 🙏

Report

1d ago

@alessandro_albolino what assurances/public auditability is there on the reliability and bias of the fact checker? Keen to understand how it remains “neutral” (assuming it does). If you allow users to customise it, wouldn’t that just lead it to become an echo chamber for existing beliefs regardless of “facts”?

Report

8h ago

Facts...No Bullsh*t

Maker

@mike_fitzbaxter The pipeline has two independent stages:

Claim detection: analyzes the transcribed audio and extracts only verifiable factual claims. It's explicitly instructed to ignore opinions, questions, and small talk. It doesn't decide what's true or false — it just identifies what can be checked.
Fact-checking: each claim is verified via live web search, not from a static knowledge base. It returns a verdict (TRUE / FALSE / UNCERTAIN), a brief explanation, and a direct link to the source so you can verify it yourself.
On neutrality and bias:
- There's no user customization of verdicts. Everyone gets the same result for the same claim — there's no "check only things I disagree with" option, so no echo chamber by design.
- The tool doesn't hold positions on topics. It delegates verification entirely to external web sources.
- The UNCERTAIN verdict exists specifically for cases where evidence is conflicting or insufficient. It doesn't force a binary true/false when reality is more nuanced.
Honest limitations:
- Bias can enter at two levels: what the llm considers a "verifiable claim" and which sources the llm surfaces from the web. No fact-checking system, human or automated, is fully immune to this.
- The quality of results depends on the quality and availability of sources on that specific topic.

Report

6h ago

Congratulations on the launch! Does the user have any control over which sources are used for verification? Also my speech was automatically translated into Italian for the transcript, claims, etc.. It still correctly detected a mixture of true and false statements, but people will almost certainly want to see the transcript in the same language as the incoming audio.

Report

9h ago

Facts...No Bullsh*t

Maker

@hex_miller_bakewell Hi, I have fixed the language detection. Now it s in the same language of the audio. Thanks for the feedback

Report

6h ago

@mike_fitzbaxter The pipeline has two independent stages:

Claim detection: analyzes the transcribed audio and extracts only verifiable factual claims. It's explicitly instructed to ignore opinions, questions, and small talk. It doesn't decide what's true or false — it just identifies what can be checked.
Fact-checking: each claim is verified via live web search, not from a static knowledge base. It returns a verdict (TRUE / FALSE / UNCERTAIN), a brief explanation, and a direct link to the source so you can verify it yourself.
On neutrality and bias:
- There's no user customization of verdicts. Everyone gets the same result for the same claim — there's no "check only things I disagree with" option, so no echo chamber by design.
- The tool doesn't hold positions on topics. It delegates verification entirely to external web sources.
- The UNCERTAIN verdict exists specifically for cases where evidence is conflicting or insufficient. It doesn't force a binary true/false when reality is more nuanced.
Honest limitations:
- Bias can enter at two levels: what the llm considers a "verifiable claim" and which sources the llm surfaces from the web. No fact-checking system, human or automated, is fully immune to this.
- The quality of results depends on the quality and availability of sources on that specific topic.