DOCUMENTATION

Verify, install, and complete the first local generation

Complete installation commands and first-run steps live here. The download page selects a platform; this page executes the flow.

DOCUMENTATION

Step one: verify the installer

Download the installer and matching SHA256SUMS file only from the official GitHub Release. Stop and download again if the digest differs.

macOS Terminal

shasum -a 256 Image-Studio-0.2.2-macos-arm64.dmg

Windows PowerShell

Get-FileHash .\Image.Studio_0.2.2_x64-setup.exe -Algorithm SHA256

macOS installation

Apple Silicon · macOS 11+

  1. 1

    Get the files

    Download the macOS DMG and SHA256SUMS-macos.txt.

  2. 2

    Verify the digest

    Run shasum in Terminal and confirm an exact match.

  3. 3

    Open the disk image

    Open the DMG and confirm it contains Image Studio.app and an Applications shortcut.

  4. 4

    Drag to install

    Drag Image Studio.app into Applications, then launch it.

  5. 5

    Handle Gatekeeper

    Only after a matching digest, remove quarantine from that app if macOS reports it as damaged.

Handle Gatekeeper only after the digest matches

sudo xattr -rd com.apple.quarantine "/Applications/Image Studio.app"

Windows installation

Windows 10 22H2 / 11 · x64

  1. 1

    Get the files

    Download the EXE and SHA256SUMS-windows.txt.

  2. 2

    Verify the digest

    Use PowerShell Get-FileHash and compare the result.

  3. 3

    Run the installer

    Run the current-user NSIS package with its offline WebView2 component.

  4. 4

    Handle SmartScreen

    Handle the unknown-publisher prompt only when the digest matches.

DOCUMENTATION

First launch and model preparation

  1. 01 Launch the app and let the device probe identify the GPU and available runtimes.
  2. 02 Open Models, choose a model that fits the current tier, and prepare its weights.
  3. 03 Reserve space for the model download and preparation cache, then wait for completion.
  4. 04 Return to Create, write a prompt, and send the first local job.

Model weights and private runtimes are separate downloads; first-time preparation needs a network while generation runs locally.

Local data and network boundary

Conversations, works, jobs, and settings are managed by local app storage and SQLite. Model acquisition, release checks, and external help links use the network.

DOCUMENTATION

Troubleshooting and support

Confirm version, platform, GPU, model, and error details, then open a reproducible GitHub Issue. The site does not provide live support.

pinned release notes