lucascs@proton.me Lucas C.S. OpenAI Whisper speech recognition models converted to the ggml format read by whisper.cpp, pinned to one Hugging Face commit and installed into /usr/share/whisper-models/ as ggml-MODEL.bin. Zeo's voice input (app-editors/zeo, USE voice) reads them from there before it offers to download a model. Install ggml-tiny.bin (78 MB): the fastest model, the least accurate Install ggml-base.bin (148 MB): the model app-editors/zeo uses by default Install ggml-small.bin (488 MB): slower, more accurate ggml-org/whisper.cpp