Rebutl
Never lose a debate to a contradiction you didn't catch.
Rebutl listens to live debates, reactions, interviews and podcasts. It separates your voice from everyone else's and feeds a private heads-up display with contradictions, evasions, position shifts and factual errors as they happen, each one grounded in evidence.
Now: “I never supported it.” At 00:12:04: “I backed it from day one.”
Question on funding went unanswered (2nd time)
“40% of…” Source shows 14%
Live conversation moves faster than anyone can fact-check. A guest contradicts what they said twenty minutes ago, dodges a question, or cites a number that's wrong, and the moment passes. Creators and debaters re-watch hours of footage to find what they should have caught live.
From first step to done.
Prepare in the War Room
Research the opponent, map their past positions, and build your case.
Go live
Rebutl separates your mic from system audio, identifies speakers, and transcribes in real time.
Read the HUD
Contradictions, evasions, shifts and factual errors appear privately, with the evidence attached.
Mine the replay
The Media Engine turns long recordings into clip cards and scripts.
What Rebutl does.
Live HUD
A private overlay of flagged moments during the conversation.
Speaker separation
You on one channel, everyone else on another, with speakers identified.
Contradiction detection
Catches statements that conflict with earlier ones, in the same session or across past ones.
Evasion and shift tracking
Notices when a question goes unanswered or a position moves.
Fact checks
Claims checked against evidence, with sources.
War Room
Pre-session research and opponent mapping.
Media Engine
Three-hour videos turned into clip cards and scripts.
Inference is labeled as inference.
Every flag Rebutl raises comes with the quote and timestamp that triggered it. Behavioral observations are clearly labeled as inference, not proof, so you always know what you can and can't say on air.
- Quote and timestamp on every flag
- Clear inference-versus-fact labels
- Headphones-only live mode
A research team in your ear.
Rebutl runs specialized engines in parallel: transcription, speaker separation, claim extraction and contradiction search, all feeding one calm display.
- Real-time transcription
- Claim extraction
- Cross-session memory
Built for
Trust & security
- Private HUD, never broadcast
- Local-first session storage
Works with the portfolio.
Real-time intelligence for live conversation.
Rebutl is in development. Ask to hear when it's ready.