AI INTERVIEW DISCLOSURE
This disclosure was prepared without legal-counsel review (none was
available). It is provided in good faith; have a qualified lawyer review it
before relying on it.
This page explains what Encase captures, processes, stores, and discards when
you take a practice interview.
Last Revised: 2026-07-17
1. WHO WE ARE
Green Letter Technologies Pvt. Ltd. (operator of Encase) provides an AI
mock-interview practice tool. This page explains what we capture, process,
store, and discard when you take a practice interview.
2. THIS IS AN AI SYSTEM (TRANSPARENCY)
Your interviewer is an AI avatar / AI voice agent, not a human. You are
conversing with an automated system that uses speech-to-text, a large
language model, and text-to-speech, with an optional animated avatar.
3. PURPOSE & NON-USE
Feedback is generated solely for your own self-improvement.
It is
never shared with employers or recruiters and is never used to make or
inform any hiring, employment, or candidate-selection decision.
(If this product were ever offered to or used by employers to evaluate
candidates, it could fall under the EU AI Act's high-risk recruitment
regime and would require re-assessment before such use.)
4. WHAT WE CAPTURE & HOW WE PROCESS IT
- Audio of your spoken answers → converted to a text transcript (Deepgram). The transcript is retained; the audio is retained only briefly, then deleted — including if the session fails, times out, or is abandoned (see §6).
- Transcript analysis by an AI language model for grammar, structure, vocabulary, and answer quality.
- Audio-intelligence over your speech (sentiment, filler words, disfluencies, speaking pace).
- Facial metrics computed on your own device (in your browser): geometric measures only — eye-contact direction and head pose. We do not infer emotions, mood, or affect. Your webcam video is processed locally and is never uploaded — only a small numeric summary of these geometric measures is sent to us.
5. FACIAL-METRICS TRANSPARENCY (NOT EMOTION RECOGNITION)
Our client-side facial metrics are purely geometric (gaze direction, head
pose). They are
not emotion-recognition: we do not classify your emotional
state. None of these metrics are used for any employment decision (see §3).
(If we ever add affect/emotion inference, we will update this disclosure
and provide the corresponding EU AI Act emotion-recognition transparency
notice.)
6. WHAT IS STORED VS DISCARDED
Stored long-term: transcript, derived scores/feedback, the geometric
facial-metrics summary, consent records. Not stored long-term: raw webcam
video (never uploaded); raw audio, which is deleted as soon as
post-session audio analysis succeeds and, in every case — including a
failed, timed-out, or abandoned session — is automatically deleted by a
storage lifecycle policy within 24 hours.
7. THIRD-PARTY PROCESSORS
Deepgram (speech-to-text + audio-intelligence), Pipecat Cloud + Daily
(real-time pipeline/transport), Simli (avatar), an Admin-configured
large-language-model provider (currently Google Gemini and/or models via
OpenRouter), and Google Cloud Platform (hosting, storage, secrets).
8. CROSS-BORDER TRANSFERS
Processing may occur outside India by the service providers above (including
in the United States) under contractual safeguards consistent with the DPDP
Act and, where applicable, GDPR standard contractual clauses.
9. YOUR RIGHTS
Under India's DPDP Act 2023 (and, where applicable, GDPR / the EU AI Act)
you may access, correct, or request deletion of your data and withdraw
consent. We process this data based on your consent captured before each
session. To exercise your rights, contact [email protected].
10. RETENTION
Transcripts, reports, and consent records are retained until you delete your
account or request deletion. Raw media is not retained (see §6).
11. CHANGES & CONTACT
We may update this disclosure; the version applicable to your session is
recorded with your consent. Questions:
[email protected].