Fix: Gemini Live model → gemini-3.1-flash-live-preview, robust JSON parsing for exam generation

Fix Gemini Live: update model to gemini-2.5-flash-preview-native-audio-dialog, fix audio extraction from model_turn, add transcription config

Add openai SDK for OpenRouter, remove anthropic (using OR instead)

Route ALL chat models through OpenRouter (free tier), add openai SDK, 6 models available

Replace google-generativeai with google-genai in requirements

Fix critical bugs: move provider to InferenceClient constructor, migrate to google-genai SDK

Update README with Gemini Live documentation

Add ffmpeg and libsndfile for audio processing

Add google-genai, pydub, librosa for Gemini Live support

Fix Gradio component wiring for Viz/Exam tabs, clean up full app

Add Gemini Live real-time bidirectional voice tab with native audio I/O

Add pandas to requirements for visualization support

Fix exec builtins for visualization code execution, add pandas support

Improve robustness: handle DeepSeek think tags, better error handling, cleaner UI

Remove gradio from requirements (handled by sdk_version)

Pin Gradio to 5.43.1 for stable API compatibility

Add README with Space metadata and documentation

Add requirements.txt

Add packages.txt for system dependencies (tesseract OCR)

Add comprehensive Study Assistant app with 6 tabs

initial commit