About
Verankila is an independent project built by Samuel Harrold, based in Indianapolis, Indiana. It was previously developed under Synavistra; as of July 2026, Verankila is owned entirely by Samuel Harrold personally and is no longer a Synavistra product.
The name
Verankila is inspired by words for root, noticing and rhythm (Malayalam veru, Japanese ki, Sanskrit laya) — for noticing where an expression starts, and how it comes and goes.
Why this exists
The scientific basis for inferring emotion from a face is genuinely contested — and Verankila is built around that fact rather than ignoring it. Every reading is labeled a tentative cue, never a clinical assessment or a fact. The point isn't certainty; it's a prompt to notice and reflect on your own expression, with a transparent explanation of what the model actually read.
Privacy by construction
All processing happens on-device. No camera frame, facial landmark, blendshape value, or emotion reading is ever transmitted off your device — see Privacy for the specifics.
Credits
Verankila's expression model was trained on facial-movement measurements derived from the Padova Emotional Dataset of Facial Expressions (PEDFE), used under its CC0 dedication: Miolla, Cardaioli & Scarpazza (2023), “Padova Emotional Dataset of Facial Expressions (PEDFE): A unique dataset of genuine and posed emotional facial expressions,” Behavior Research Methods, 55(5), 2559-2574, doi.org/10.3758/s13428-022-01914-4. No PEDFE image or video is included in the app.
Support
Something not working, or feedback on how Verankila behaves? Email support@verankila.app — every message is read.