You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
An adaptive, voice-driven mock-interview platform for **Data Scientist**, **Algorithm
6
-
Researcher**, and **AI Engineer** roles: a lip-synced 3D interviewer asks questions planned
6
+
Researcher**, and **AI Engineer** roles: an animated 3D interviewer asks questions planned
7
7
for your target role and difficulty, listens to your spoken answers, asks follow-ups, offers
8
8
hints, scores every answer on an 8-metric rubric, and produces a full readiness report with a
9
9
study curriculum that tracks your progress across sessions.
@@ -30,7 +30,7 @@ working fallback** — the app runs fully offline with no API key.
30
30
31
31

32
32
33
-
*A live interview in **English**: the interviewer speaks the question with a lip-synced
33
+
*A live interview in **English**: the interviewer speaks the question with an animated 3D
34
34
character, and you can **barge in** — start answering mid-question and it stops to listen
35
35
(an echo guard keeps it from hearing its own voice).*
36
36
@@ -196,7 +196,7 @@ By default the interviewer speaks with the browser's `speechSynthesis` and an an
196
196
avatar — no setup needed. For the full experience, a local **HeadTTS** sidecar (MIT) serves
197
197
**Kokoro-82M** (Apache-2.0) TTS on port 8012, fully offline after a one-time model download.
198
198
Each utterance returns word + viseme timelines, so the **real-time 3D talking head**
199
-
(met4citizen/TalkingHead, Three.js) lip-syncs exactly to the audio from the first frame.
199
+
(met4citizen/TalkingHead, Three.js) animates the mouth in exact sync with the audio from the first frame.
200
200
201
201
```powershell
202
202
powershell -ExecutionPolicy Bypass -File scripts\setup_voice.ps1 # Windows provisioning
0 commit comments