Detailed sentiment analysis for spoken video

Upload a video of someone speaking. Spectra transcribes it, then nine machine-learning models read the words, the voice and the face to score 40+ categories — from misogyny, antisemitism and Islamophobia to sarcasm, propaganda and empathy. You'll receive a private link to your report by email.

Speakers are separated automatically; pick a number if you know it for better accuracy. Language is set to Urdu by default (automatic detection often mistakes Urdu for Hindi); change it for other videos. Thorough runs every category on every segment. Fast screens segments first and runs the detailed checks only where something looks off.

Ordered on Fiverr or Upwork? Enter your order number instead of an email.

Thank you, your file arrived

Upload another