LattifAI Python v1.0.0 is a significant update that introduces a refactored caption system, enhanced alignment engine, and improved audio processing capabilities.

✨ Major Changes

🔄 CLI Architecture Refactor

New Command Structure

Migrated to a unified subcommand-based CLI powered by nemo_run (lattifai-run) framework:

📝 Caption System Refactor

New Unified Caption Class

Introduced a comprehensive Caption class that replaces the previous CaptionIO for streamlined subtitle/caption handling:

from lattifai.caption import Caption

# Read captions with metadata preservation
caption = Caption.read("subtitles.srt")
print(f"Loaded {len(caption)} segments")
print(f"Language: {caption.language}")
print(f"Duration: {caption.duration:.2f}s")

# Write to different formats
caption.write("output.vtt")
caption.write("output.TextGrid")

🌐 Multilingual Tokenization

Added text tokenization function supporting multiple languages:

# Supports Chinese, English, and German text processing
# Improved handling of mixed-language content