Open-source software and models

QuietScribe uses software and models from the projects below. Their respective authors retain ownership. The licenses listed here govern those components; QuietScribe’s commercial terms do not restrict the rights they grant.

FFmpeg source and build instructions

This software uses FFmpeg under LGPL 2.1 or later. QuietScribe runs FFmpeg and FFprobe as separate executables. Download the corresponding source, license and build instructions here:

• ffmpeg-8.0-6a392b02f7b273c0: source archive · build script · instructions · checksums

Notices

• whisper.cpp — MIT license.
• FFmpeg — LGPL 2.1 or later.
• sherpa-onnx — Apache 2.0.
• ONNX Runtime — MIT license.
• Ollama, downloaded during setup — MIT license.
• Whisper models, downloaded during setup — MIT license.
• Gemma 4, downloaded via Ollama — Apache 2.0 and attribution.
• Pyannote speaker segmentation, downloaded during setup — MIT license and attribution.
• NVIDIA NeMo TitaNet-S, ONNX export distributed by sherpa-onnx — model attribution and Apache 2.0 notice.

Notices are also included inside QuietScribe.app at Contents/Resources/engine/licenses. Names and logos of third-party projects belong to their owners; listing them does not imply endorsement.

Contact info@digitalitup.in if you cannot access the source or need help identifying the source for a particular release.