Last updated September 14, 2026

Third-Party Notices

Speech Connect includes or can download third-party software, models, and libraries that run locally as part of the desktop app. This page summarizes major components used by current builds; release-specific license files or links should be reviewed with each public installer.

Major components

• Electron, Chromium, Node.js, React, Vite, TypeScript, and related JavaScript packages.
• ffmpeg for audio conversion.
• whisper.cpp and Whisper-compatible speech recognition model files.
• llama.cpp and local language/embedding model files.
• Kokoro voice models and related runtime dependencies.
• Expressive local voice lab dependencies, including Chatterbox-related Python packages when installed by the app.
• SQLite and optional vector search components.

Model notices

Model files can have their own licenses, acceptable-use terms, and attribution requirements. The app may download local model assets, but Adeptimus does not receive your app content when those local models are used.

Release-specific notices

Public installers should include the actual third-party notices, license texts, and attribution links for the exact components and model assets packaged with or downloaded by that release.