Spaces:
Sleeping
Sleeping
Commit History
-
deploy: /admin/ai/keys — report which keys are in rotation
7960c35 -
deploy: alert at 40/50/60/70/80/90/95/100% of each key's daily budget
3443a75 -
deploy: rotate across all Groq keys, alert admins at 80/100%
edc0702
-
Restore AI: Groq retired the Llama models
2f666f7
-
deploy: Groq failures become clear HTTP errors, not bare 500s
A hit token cap now returns a generic 'AI service is busy, try again' instead of 'Internal server error'; the real quota reason stays in the server log only.96abfbd
-
feat(mock-interview): Whisper speech-to-text for spoken answers
Sync from GitHub feat/mock-interview-upgrades (b3b5a45): adds POST /mock-interview/transcribe (Whisper large-v3 via Groq, English forced, question text as vocabulary hint, 503 on failure so the client falls back to the browser transcript). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>e0567f6 -
Deploy mock-interview upgrades + streaming TTS proxy
Syncs the Space with the GitHub feat/mock-interview-upgrades branch: interview length tiers, resume/JD tailoring, adaptive follow-ups, model answers, delivery scoring, progress endpoint, seniority/focus knobs, zod-validated LLM output — and the /mock-interview/tts proxy, which now pipes audio streamed from the Python (Kokoro) service and forwards voice/format. New entity columns are nullable/defaulted, so TypeORM synchronize adds them to the existing DB without breaking rows. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>0d43de6
-
feat: add mock interview and document generation modules
Hajrah Rehan9ffaa46
-
new commit by Mahroosh
mahrooshishaq9a951d7