InvokeAI Database Tools

A collection of utility scripts for working with the InvokeAI SQLite database.
These tools are designed for recovering lost image entries, rebuilding the image index, and reclassifying images as user assets and vice versa.

⚠️ WARNING: ⚠️

Always back up your invokeai.db before running any scripts!

These tools modify the database directly and changes cannot be undone.

---

Features

Recover missing images from <OUTPUTS_PATH>/images/

Script: Restore_Images_DB_v2.1.py

This script scans the directory you provide via --outputs and compares files to the entries stored in the database.
Any images that exist on disk but are missing from the database are added back. The path is defined by the string in your invokeai.yaml. Script also attempts to recover any metadata image might have.

The script:

• Scans <OUTPUTS_PATH>/images/
• Detects all PNG images not present in images table of cuirrent DB
• Inserts them with:
  • image_category = "general"
  • image_origin = "internal"
  • is_intermediate = 0
• Creates a board named Recovered DD-MM-YY
• Adds all recovered images to that board

---

Convert all images from defined board into user assets keeping the board assignment

Script: Convert_Board_to_Assets.py

This tool updates images belonging to a specific board and classifies them as assets.

It:

• Locates a board by name (Not case-sensitive)
• Reads all image_name entries from board_images
• Updates the corresponding rows in the images table:
  • image_category = "user"
  • image_origin = "external"
• Does not modify board membership or any other fields

---

Convert all images from a defined board into regular gallery images keeping the board assignment

Script: Convert_Assets_to_Board_v1.0.py

This tool updates images belonging to a specific board and classifies them as regular gallery images.

It:

• Locates a board by name (not case-sensitive)
• Reads all image_name entries from board_images
• Updates the corresponding rows in the images table:
  • image_category = "general"
  • image_origin = "internal"
• Does not modify board membership or any other fields

Useful for reorganizing imports or turning grouped images into asset references.

---

1. For users launching InvokeAI via the official Launcher (recommended)

Running the scripts using the InvokeAI Launcher

If you use the official InvokeAI Launcher, you don’t need to activate the virtual environment manually. The Launcher provides a dedicated Dev Console with the venv already activated. Just click >_ button in lower left corner

Steps

1. Open the InvokeAI Launcher.

2. Click >_ button in lower left corner. Then Start Dev Console This opens a terminal window with the proper virtual environment activated.

3. In the Dev Console, run:

   python path/to/Restore_Images_DB_v2.1.py --db path/to/invokeai.db --outputs path/to/outputs

   or

   python path/to/Convert_Board_to_Assets.py --db path/to/invokeai.db --board-name "Board name to convert" --verbose

   or

   python path/to/Convert_Assets_to_Board_v1.0.py --db path/to/invokeai.db --board-name "Board name to convert" --verbose

---

2. Running the scripts using InvokeAI’s virtual environment in terminal of your choice:

Launch it manually via PowerShell or CMD, you can use the same virtual environment as InvokeAI to run these tools. No separate Python installation needed.

Windows (CMD or PowerShell)

1. Open a terminal.

2. Navigate to your InvokeAI installation directory:

   cd path\to\InvokeAI

3. Activate the virtual environment:

   .venv\Scripts\activate

4. Run the script:

   python path\to\Restore_Images_DB_v2.1.py --db path\to\invokeai.db --outputs path/to/outputs

   or

   python path\to\Convert_Board_to_Assets.py --db path\to\invokeai.db --board-name "Board name to convert" --verbose

   or

   python path\to\Convert_Assets_to_Board_v1.0.py --db path\to\invokeai.db --board-name "Board name to convert" --verbose

Linux / macOS

1. Open a terminal.

2. Navigate to your InvokeAI directory:

   cd /path/to/InvokeAI

3. Activate the venv:

   source .venv/bin/activate

4. Run the script:

   python /path/to/Restore_Images_DB_v2.1.py --db /path/to/invokeai.db --outputs /path/to/outputs

   or

   python path/to/Convert_Board_to_Assets.py --db path/to/invokeai.db --board-name "Board name to convert" --verbose

   or

   python path/to/Convert_Assets_to_Board_v1.0.py --db path/to/invokeai.db --board-name "Board name to convert" --verbose

   This uses the same Python environment as InvokeAI, so no extra dependencies are required.

---