First hour with your AI image generator
You finished the Stability Matrix install. This page is the first-hour checklist for a local AI image generator: confirm the launcher, pick a data directory, install one package, and keep one update door. Home install hub: AI image generator.
1. Launch and confirm the shell
Open Stability Matrix from your extract folder, Applications folder, or AppImage path. You should see the package manager UI. If nothing opens, re-check the asset name from GitHub Releases before you assume a broken AI image generator binary.
2. Confirm the data directory
Accept or choose the portable data folder so packages and models stay with this install. Write the path down. Shared PCs should avoid other users’ private folders for the AI image generator library.
3. Install one package only
Pick a single supported package from the one-click list. Wait for the install to finish. Do not queue three stacks on minute one; the AI image generator launcher already gives you a path to expand later.
4. Launch that package once
Start the package from the launcher and confirm the UI loads. That single proof teaches the local AI image generator contract better than a feature video. If launch fails, read the package log before you retry with larger models.
5. Import or download one small model
Use the model browser or drag a known checkpoint into the shared folder. Prefer a small file for the first test so disk and network issues show up early.
6. Optional Inference workspace
Open Inference tabs only after a package works. The built-in workspace helps when you want a local prompt UI without a browser SaaS seat. Skip it if your chosen package already covers the job.
7. Keep one update door
Write down whether you installed from Windows zip, macOS DMG, or Linux AppImage. Return to that same door for upgrades. Releases overview: releases. Safe habits: download safe.
8. Schedule a real project folder
Pick one real output folder for tomorrow’s generations. Use the AI image generator with models you trust. If you need to leave browser-only habits behind more formally, read local vs cloud.
Troubleshooting: troubleshooting. Safety: is it safe. First run notes: first run. Comparison context: vs browser AI tools.
Blocked first launches
If SmartScreen or Gatekeeper blocked the first launch, reopen the original file from GitHub Releases rather than a second website. Filename tokens such as win-x64 should still match. Portable users should confirm they launched the binary inside their extract folder, not a partial download.
Onboarding checklists
Teams can turn this hour into a short checklist in onboarding docs: launch, data directory, one package, one model, one update door. That keeps the AI image generator useful after reimages without rediscovering adware SERPs.
Power features wait
Power features such as training UIs, extra packages, and custom launch args are upstream strengths. They are not the first-hour bar. Prove the AI image generator with one boring package first, then grow into advanced menus with less risk.
Finish line
When you finish the hour, you should be able to say which asset you installed, where the data directory lives, and which package you launched. Those three facts make support tickets shorter and upgrades safer.
Focus and fonts
Enlarge UI fonts until settings remain readable on your display. Resist installing every package before two successful launches. Both habits reduce accidental clicks on the wrong download later.
Record the asset name
Write the asset name into your notes: StabilityMatrix-win-x64.zip, macos-arm64.dmg, or linux-x64.zip. That single line prevents next month’s “which binary is this” thread when someone asks how this AI image generator was installed.
A second proof after lunch helps memory: open the launcher again, confirm the package, and run another tiny generation. That repetition beats reading a feature list. Keep Releases bookmarked so the next upgrade does not start with a search ad.
Shared PC notes
On a shared workstation, write the AI image generator asset name on a sticker note or wiki line. The next person should not have to guess whether the machine used the Windows zip, the macOS DMG, or the Linux AppImage. Clear ownership of the data directory prevents orphaned multi-gigabyte folders.
If antivirus quarantines a fresh extract, compare the filename to GitHub Releases before you restore anything. A drifted name is a delete-and-redownload event, not a reason to trust a random mirror for the AI image generator launcher.