CSCSoftware/vibeaigent-models

Vibe Aigent model mirror

Pinned copies of the model files that the Vibe Aigent setup wizard downloads for its optional local voice features. The files are unchanged copies of the upstream releases listed below; the wizard verifies every download with the SHA-256 shown here.

File	Purpose	Upstream	License	Size	SHA-256
whisper/ggml-large-v3-turbo-q5_0.bin	Local speech recognition (whisper.cpp / Whisper.net)	ggerganov/whisper.cpp (OpenAI Whisper large-v3-turbo)	MIT	574,041,195	394221709cd5ad1f40c46e6031ca61bce88931e6e088c188294c6d5a55ffa7e2
qwen3-tts/qwen-talker-1.7b-base-Q8_0.gguf	Local announcement voice (qwentts.cpp)	Serveurperso/Qwen3-TTS-GGUF (Qwen3-TTS by the Alibaba Qwen team)	Apache-2.0	2,079,448,256	4b9a33a236908dd9435a42f7a396e38038329d053b704342a6413c08544c4fda
qwen3-tts/qwen-tokenizer-12hz-Q8_0.gguf	Audio codec for the voice above	Serveurperso/Qwen3-TTS-GGUF	Apache-2.0	291,150,624	1883beeed99348fc35e23dd225e9082f93f6f8c109330a33d935baa8acdbfd94

| voices/voices.zip | 14 voices (.qvoice): 10 German (5 male / 5 female, also speak English) and 4 native English (2 / 2) for the local announcement voice | CSC-Software GmbH (created and re-trained) | Vibe Aigent voice pack license (LICENSE-voices.txt inside: free use incl. commercial, no impersonation) | 33,742,254 | 3b4cbb18677128c964ea48c6cc30bc2915356bc83b879422745bb475f2c7886f |

All credit for the models goes to their authors (OpenAI Whisper and the whisper.cpp project; the Qwen team at Alibaba and the qwentts.cpp project). The license of each file is the upstream license named in the table.

Programs (Vibe Aigent, the qwentts.cpp server build) are not hosted here; they come from GitHub releases.

Publisher: CSC-Software GmbH · https://www.csc-software.de