Your Eyes Are Lying to You. I Can Prove It.
Relying on intuition to spot a deepfake in 2026 is like bringing a knife to a drone fight. Fake or manipulative content has become a scourge. Between high-fidelity voice clones, synthetic manipulation, and influence actors with media access, “seeing is believing” has become a dangerous liability.
If you work in a newsroom, a law firm, or a research lab, your gut feeling no longer counts as a forensic tool. I’m so excited to share that my new app Kinexis.AI goes live today to help bridge the gap between human perception and manipulated reality.
Instead of a simple “fake or not” detector, my company has built a multi-modal telemetry engine. Running a video through Kinexis generates a clinical, forensic report across multiple layers of analysis:
Synthetic Artifacts: Hunting for the microscopic glitches GenAI leaves in its wake.
Narrative Framing: Identifying hidden persuasion techniques like authority positioning or “feigned earnestness.”
Body Language: Mapping movements and expressions that often occur too fast for the human eye to register.
Jeffrey Epstein’s deposition footage serves as a quick case study.
When we processed clips of those archival tapes, the engine didn’t just confirm they were authentic; it flagged specific moments of contempt and high cognitive load that might change your understanding of the testimony.
Kinexis turns a passive viewing experience into a data-driven investigation.
Building Kinexis has been a deeply personal mission over the past weeks because I believe technology should act as a shield for truth rather than as a weapon against it.
This platform is a first-of-its-kind attempt to complement every citizen and professional’s valuable human judgment with a high-fidelity AI sensor that offers a nuanced, forensic understanding of how media can be constructed to influence us.
By merging deepfake detection with behavioral intelligence, my hope is that we can use AI for good—empowering people to cut through the noise of synthetic manipulation and restore a shared, factual foundation for how we understand the world.
I built this for those who hate black-box verdicts, but haven’t had useful tools to dig deeper. You’ll get raw data, timestamped findings, and probabilistic confidence scores.
Stop guessing. Start measuring!
You can run your first analysis right now, or browse public reports. Guest checkout is enabled, and we support direct links from sites like YouTube, Facebook, X/Twitter, and TikTok so you don’t have to waste your time downloading videos.
Stay skeptical.
This is an incredible idea and so desperately needed. I look forward to checking it out today!
I'm sure you all likely already thought of this but, if not yet, pitching this to services like Ground News, Politifact and other fact checking services, the AP, and independent journalists would likely be well received. I'll be sure to share it wherever I can.
As an aside, my son's good friend, who's he's attending college with, has been working on creating a very similar AI application for this type of usage for the last few years. I believe his has some differences in how it goes about what it does, but the psychological evaluation and assessment of content is similar. I'll have to let him know to check out your app and that you may have beat him to the finish line!
Thank you for doing this. Time seems be exceedingly short.
How can I send you non-publicly Claude’s PDF of our convo today on “Preserving Democratic Autonomy in the USA” full conversation thread • Claude.ai • March 10, 2026 ?