Voicitool

AI disclosure: The code of this tool was written with help from Claude, Anthropic's AI assistant (https://claude.ai). I tested everything and the full source code is on GitHub. The tool itself uses AI models to separate voices, recognize speech and tell speakers apart (see Credits). It does not clone or generate any voices. Voicitool makes dub packs for The Choicer Voicer out of any video. Drop in a clip or paste a YouTube link. Voicitool pulls the voices away from the music, writes down what's said, works out who says what and cuts every line into its own clip. Then you go through it in the editor, fix what it got wrong, and export the pack or put it straight into the game. Everything runs on your own PC. Features Separates voices from music and background Speech recognition in about 100 languages Finds the speakers, you just give them names Editor with waveform, keyboard shortcuts and undo Finds laughs and adds them as lines Export as a ZIP or install straight into the game Categories for bigger collections, like 30 Family Guy clips Comes in 26 languages and updates itself Install Download Voicitool.exe from GitHub: https://github.com/Jan1653/Voicitool/releases/latest Run it. Windows may warn you because the exe isn't signed. Click More info, then Run anyway. The setup downloads everything it needs, about 10 GB. That takes 10 to 40 minutes. Requirements Windows 10 or 11, 64-bit 8 GB RAM (16 GB recommended) and about 15 GB of free space An NVIDIA graphics card is recommended. Without one it still works, it just takes a lot longer. About the virus warning Windows Defender sometimes flags the exe as "Wacatac.B!ml". That's a guess by its automatic detection, and it hits a lot of new tools that aren't signed. It has been reported to Microsoft. All the code is on GitHub if you want to check it or build the exe yourself. Source code and bug reports: https://github.com/Jan1653/Voicitool Credits Code written with the help of Claude by Anthropic: https://claude.a