🔒 1. Local-Only Execution Architecture
All transcription, audio dead-air removal, motion typography animation, and video rendering execute 100% offline on your personal machine. No voice recordings, transcribed text, background audio, or rendered MP4 files are ever uploaded or transmitted to any external server.
🛡️ 2. Automatic 100% Metadata Stripping
Whenever a video is produced, the engine runs an automated metadata purification pass using FFmpeg's bitexact sanitizer. This strips camera IDs, hardware identifiers, author profiles, software build markers, GPS tags, and file creation timestamps from the container.
🎧 3. Auto-Deletion of Source Audio
To prevent storage clutter and protect unreleased spoken thoughts, the pipeline automatically purges the raw source audio file from the Audio/ directory immediately after successful video generation.
🌐 4. Self-Hosted Fonts & Zero Third-Party CDNs
All typography fonts (JetBrains Mono, Plus Jakarta Sans, Space Grotesk) are self-hosted directly within the local repository. When you view this site or use the software, no web requests are made to Google Fonts servers, guaranteeing your IP address is never logged.
🤖 5. Local AI & LLM Processing
If you enable local AI styling via Ollama (e.g., qwen3.5:9b or llama3.1), the model operates strictly on http://127.0.0.1:11434. No prompts or transcripts leave your computer.
📜 6. Open Source Transparency (MIT License)
typo Videos is free and open source. You can inspect every line of Python, batch scripting, and CSS in the repository to verify that zero tracking or surveillance code exists.