VoiceStudio
50,000Download

v0.5.6 · released Sep 23, 2026 · no account required

Desktop builds, verified byte-for-byte.

macOS, Windows, and Linux. No account required. Every artifact is checksum-published and signature-linked before it gets a button here.

50,000
GitHub stars
~596,000
Downloads
103,171+
Docker pulls

Install VoiceStudio · Free · Open source

curl -fsSL https://voicestudio.sh/install | sh
  • v0.5.6

    macOS (Apple Silicon)

    M1 · M2 · M3 · M4 · 255 MB

    Download
    Verify download
    Architecture
    arm64 / aarch64
    Requires
    macOS 13.3 or newer on Apple Silicon
    File
    VoiceStudio-Electron-0.5.6-mac-arm64.dmg
    SHA-256
    821d011faf255cd6315dba969ce1572ef850ae6159fcc55d94dcb9fe86259834
    Signing
    Ad-hoc signed; the published DMG is not notarized
    Verification instructionsPublished checksums
  • v0.5.6

    macOS (Intel)

    Intel x64 · remote backend required · 262 MB

    Download
    Verify download
    Architecture
    x86_64
    Requires
    macOS 13.3 or newer; local Python backend is unavailable
    File
    VoiceStudio-Electron-0.5.6-mac-x64.dmg
    SHA-256
    1363e09c47225b7190b39caae5c7a6c3c3720bbd0fe7ea685c8643033f0ca25b
    Signing
    Ad-hoc signed; the published DMG is not notarized
    Verification instructionsPublished checksums
  • v0.5.6

    Windows (x64)

    Installer · 205 MB

    Download
    Verify download
    Architecture
    x86_64
    Requires
    Windows 10 or 11, 64-bit
    File
    VoiceStudio-Electron-0.5.6-win-x64.exe
    SHA-256
    3d7ac1d6b54e7854ca7fff4dda838e402d1c5d31acc9656b1838432e29a97624
    Signing
    Detached signature unavailable
    Verification instructionsPublished checksums
  • v0.5.6

    Linux (x64)

    Portable AppImage · 261 MB

    Download
    Verify download
    Architecture
    x86_64
    Requires
    Linux x86_64 with glibc 2.39 or newer
    File
    VoiceStudio-Electron-0.5.6-linux-x64.AppImage
    SHA-256
    b44d286aa862a658374846599f9692a3a1ebe56f13e4bfca6d94885aab28a90f
    Signing
    Detached signature unavailable
    Verification instructionsPublished checksums
All releases on GitHub ↗

Docker

Headless web studio — run on a GPU box, homelab, or anywhere Docker runs.

palashdeb/omnivoice-studio

FastAPI + React UI on port 3900. CUDA, ROCm, or CPU. Image on Docker Hub and GHCR.

Always the newest build: --pull always refreshes the image on every start.

Docker Hub ↗
Docker Hub
103,171+pulls
GitHub Container Registry
40,443+pulls
docker run -d --pull always --name omnivoice \
  -p 127.0.0.1:3900:3900 \
  -v omnivoice-data:/app/omnivoice_data \
  -v ~/.cache/huggingface:/root/.cache/huggingface \
  palashdeb/omnivoice-studio:latest

Open http://localhost:3900 after the container starts. First run pulls model weights. Docker install guide.

Source

Build from the repo, file issues, or contribute engines.

FAQ

Quick answers before you install.

Does it need an account or API key?

No. VoiceStudio runs entirely on your machine. Clone, dub, and dictate without signing in or talking to a cloud API.

Desktop vs Docker — which should I use?

Desktop is the full app with auto-updates. Docker is the headless web UI (port 3900) for servers, homelabs and GPU boxes. Both run the same engines.

How much hardware do I need?

About 8 GB RAM (16 GB+ recommended) and ~10 GB disk for models. GPU optional — 4 GB VRAM works with CPU offload; no GPU is fine, just slower.

Where are older versions?

Every release and asset lives on GitHub Releases. Docker has :latest for the newest build and :rocm for AMD.

What is Nightly?

Nightly builds ship from the rolling GitHub preview release (main). They get new features first and may be unstable. Use stable for day-to-day work; switch channels in Settings → About.