Control · Storage

How Exports & Backups Work

Keep your outputs and params.json in your own drive; let the container vanish.

The rule (read this first)

CloudDock containers are temporary workspaces. They are designed to be disposable: fast to start, clean to use, easy to destroy.

Non-negotiable: your data only truly belongs to you after it’s exported to your own drive (Google Drive, Dropbox, OneDrive, S3, NAS, anything you control). If the container is destroyed before export, it’s gone — and support cannot recover it.

What you should always save

  • Outputs (final images, previews, grids, samples)
  • Generation metadata (A1111 PNG info / prompts)
  • params.json (Training Center settings — the “recipe”)
  • Models you trained (LoRA / DreamBooth outputs)
Simple mindset: Treat the container like a hotel room. Enjoy it. Work in it. But don’t leave valuables in it when checkout hits.

A1111: export images the simple way

A1111 outputs are easy to keep: right-click → Save image as… and put it wherever you want.

  • Save to your computer
  • Save directly into a synced folder (Google Drive / Dropbox desktop folder)
  • Or upload to your cloud drive from the built-in browser
Right-click save image as in A1111 gallery output
A1111 outputs: right-click → Save image as → your drive of choice.

Jupyter: default workspace path

In CloudDock containers, the default working directory is: 

/workspace

If you’re using notebooks, scripts, or file-based workflows, keep your project files under /workspace and export them when done.

JupyterLab file browser showing /workspace directory
JupyterLab: start at /workspace. Keep important files organized there.

SD Training Center: where outputs and params live

CloudDock SD Training Center saves outputs to: 

/workspace/train/output

This folder typically contains:

  • Trained artifacts (LoRA / checkpoints, depending on the job)
  • Logs / progress records (if enabled by the job)
  • params.json — your training recipe
Best practice: Always export params.json. Even if you lose outputs, the recipe lets you recreate the run.
Training Center output folder path /workspace/train/output highlighted
Training Center default: /workspace/train/output — export this folder to your drive.

Upload to any cloud drive you like

Once your files are in your workspace (or downloaded to your computer), upload them to whatever storage you trust:

  • Google Drive
  • Dropbox
  • OneDrive
  • S3-compatible storage
  • Your own NAS / server
CloudDock doesn’t care where you store it. The rule is: get it out of the container.

Backup frequency: every 24 hours (minimum)

No matter what workflow you run, you should back up at least once every 24 hours. Containers are temporary by design — and “I’ll do it later” is how people lose data.

Reality check: If the container is destroyed, everything inside is gone. Support cannot recover it — not because we don’t want to, but because we literally can’t. Please budget time for backups. This is not the place to save money.

Time management: check remaining time and extend early

Backups take time — especially large training outputs. You should regularly go back to the Dashboard to check remaining time.

If you’re getting close to the end, extend the session before you start a big export.

How to extend hours

  1. Go back to the CloudDock Dashboard.
  2. Find Extend hours (or your time extension control).
  3. Enter the number of hours to add.
  4. Click the small triangle / confirm control to apply.
Dashboard showing extend hours input and small triangle confirmation control
Extend early, not last-minute. Exports are not instant.

FAQ

“Why are containers temporary?”

Disposable containers keep the environment clean, consistent, and reproducible. It prevents “state rot” — the slow accumulation of broken installs and untraceable changes.

“Can support restore my files if I forgot to export?”

No. If the container is destroyed, the data is gone. There is nothing support can pull back. Please export regularly.

“What’s the minimum backup set?”

Your final outputs + params.json. If you have time, back up the full output folder under /workspace/train/output.

What’s next?

Export first. Relax second.
← Back to SD Home