Skip to content

Audio Whisper Models not working #518

@blu3knight

Description

@blu3knight

Describe the bug
When clicking on Record in Whisper /Voice the system can not allocate size or start recording.

To Reproduce
Click on Record button in Audo Booth
Get "Speech-to-text engine not ready" popup

Model is Whisper Turbo Medium

Expected behavior
Should Work

Screenshots

Desktop (please complete the following information):

  • OS: [e.g. macOS 15, Windows 11]. - MacOS 15.7.3

  • Witsy Version [e.g. 2.2.4]

  • 3.4.2

Additional context
Logs

[2026-01-26 13:36:38.920] [info] 2026-01-26 13:36:38.923498 [W:onnxruntime:, session_state.cc:1282 VerifyEachNodeIsAssignedToAnEp] Rerunning with verbose output on a non-minimal build will show node assignments. https://cdn.jsdelivr.net/npm/@huggingface/transformers@3.7.3/dist/ort-wasm-simd-threaded.jsep.mjs:100
[2026-01-26 13:36:39.416] [info] [whisper] error when initializing: Error: Can't create a session. failed to allocate a buffer of size 1828446806. file:///Applications/Witsy.app/Contents/Resources/app.asar/.vite/renderer/main_window/assets/index-CYgvmRY2.js:12324
[2026-01-26 13:36:39.606] [info] [stt] whisper Xenova/whisper-medium file:///Applications/Witsy.app/Contents/Resources/app.asar/.vite/renderer/main_window/assets/index-CYgvmRY2.js:12774

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions