CapoScore reads face, voice, language and heart rate together, measures every sentence against the speaker’s own normal behavior, and hands you a clear, evidence-backed report.
Lift: at our review threshold, a documented false statement is ~10× more likely to be flagged than ordinary speech
Median score for documented false statements versus everything else
Documented false statements surfaced while reviewers read only ~1 in 11 sentences
U.S. patents pending covering the CapoScore methodology
* Internal evaluation, not independently audited. Person-held-out validation on our own data; independent third-party testing pending. Important disclosures.
Upload any recorded video or call — a press interview, earnings call, deposition or panel — or paste a YouTube link. There is nothing to install and nothing for the speaker to wear. Files go straight to secure, isolated storage and you can follow progress live.
Most footage has more than one person in it. CapoScore detects every face, separates who is speaking when, and links each face to its own voice — then locks onto the subject you care about, so every signal is attributed to the right person.
The same person is recognized across all of your videos, which is what makes a personal baseline possible.
For the subject alone, CapoScore reads four channels at once, moment by moment — not just averages, but the spikes, suppressions and sustained shifts in between.
Emotion and micro-expressions, frame by frame.
Tone, pitch, stress and dominance shifts in the subject’s own speech.
What is said, avoided and deflected — hedges, fillers, pauses.
A contact-free heart-rate signal read from the video itself.
Plus the signal between them: when the face says one thing and the voice another, that mismatch is evidence in its own right.
People differ — some are naturally animated, others flat. CapoScore never compares a speaker to a population average. It builds a behavioral baseline for each person and measures every sentence against their own normal, so an expressive speaker is not penalized for being expressive.
Each sentence receives a score from 0 to 100, adjusted for the setting and stakes of the conversation by our proprietary Context Risk Calculator, and placed in one of four risk tiers. Scores roll up by topic, showing the typical and the peak level for each subject discussed.
Every sentence at 65 or above gets an AI deep dive. It reads the words alongside what shifted in face, voice and heart rate, and classes the moment — with the evidence and its reasoning — before it goes to human review. One universal threshold, the same for every subject.
An interactive report built for decisions, ready to share with your team.
CapoScore shows you where to look closer. The judgment stays with you.
Browse complete reports, or talk to us about your use case.
* Performance figures. The performance figures presented on this page (including lift, median scores and detection rates) are derived solely from CapoScore Inc.’s internal, person-held-out evaluation on its own proprietary data population, consisting of 89 videos of publicly documented deception. These figures have not been independently audited, verified or replicated; independent third-party validation testing is pending. They are provided for general informational purposes only, reflect performance under specific test conditions, and may not be representative of results on other data, subjects, recording conditions or use cases. Past or tested performance is not a guarantee of future results. No representation or warranty, express or implied, is made as to the accuracy, completeness or reliability of any figure.
† Patents. Six U.S. provisional patent applications have been filed. “Patent pending” means that applications are pending; it does not mean that any patent has been granted, and there is no assurance that any patent will issue or of the scope of any claims that may be allowed.
Subject to change. All figures, methodologies, features, descriptions and other content on this page are subject to change, modification, supersession or withdrawal at any time, without notice, and CapoScore Inc. assumes no obligation to update, revise or correct this page or any information on it.
Not a determination of truthfulness. CapoScore outputs are probabilistic indicators of behavioral deviation only. They are not a determination of truthfulness, deception or intent; they are not legal, financial, employment or investigative advice; and they should never serve as the sole basis for any decision. Use of CapoScore is governed by our Terms of Service and Privacy Policy.