Yes, all Unicode characters including Chinese, Japanese, Arabic, etc.
No. All text extraction happens locally in your browser.
No. Only text-based PDFs (not scanned images) can be extracted.