Configuration Options
On this page
Configuration Editor
Tag-AI includes a graphical Configuration Editor for adjusting application settings.
To open the Configuration Editor:
- Click on Actions in the top menu bar
- Select Edit Config from the dropdown menu
Configuration Sections
The configuration file is divided into several sections, each controlling different aspects of Tag-AI.
Watch Folders
The [watch] section controls which folders Tag-AI monitors for images.
Settings
- folder: Comma-separated list of directories to scan for images
Watch Folder Management
In the Configuration Editor:
- Use the + button to add a new folder
- Click Find... to browse for a folder
- Click Remove to delete a folder from the list
Image Processing
The [image_processing] section controls how images are processed.
Settings
- target_max_size_mb: Maximum size (in MB) that images will be resized to before processing
Reducing this value:
- Speeds up processing
- Reduces memory usage
- May slightly reduce tagging accuracy
Increasing this value:
- Improves tagging accuracy for detailed images
- Increases memory usage
- Slows down processing
Gemini API Settings
The [tagger_gemini] section configures Google's Gemini API for cloud-based tagging.
Settings
- gemini_api_key: Your Google Gemini API key
- paid_version: Whether you're using the paid API tier (true/false)
- max_tags: Maximum number of tags to keep per image
- confidence_threshold: Minimum confidence score for keeping tags (0.0-1.0)
- max_image_size: Maximum dimension for images sent to Gemini API
To use the Gemini API, you must obtain an API key from Google. See the Gemini API Setup guide for details.
Ollama Settings
The [tagger_ollama] section configures the local Ollama-based tagging with LLaVA.
Settings
- ollama_endpoint: API endpoint for the Ollama service (default: http://localhost:11434/api/generate)
- model_name: Name of the Ollama model to use (default: llava)
Changing the ollama_endpoint is only recommended for advanced users who have configured a custom Ollama setup.
Script Selection
The [script_to_run] section determines which AI model is used for tagging.
Settings
- llm_script: Which AI model to use
local: Use local Ollama/LLaVA processing (default)gemini: Use Google's Gemini API for cloud processing
License Settings
The [license] section contains your Tag-AI license information.
Settings
- license_key: Your Tag-AI license key
Do not modify the license key manually. The correct key is set during activation.
Setup Status
The [setup] section tracks whether initial setup has been completed.
Settings
- setup_complete: Whether the setup wizard has been completed (true/false)
- gpu_type: Detected GPU type
- gpu_acceleration: Whether GPU acceleration is enabled
- gpu_setup_skipped: Whether GPU setup was skipped during installation
Manually changing setup_complete to false will trigger the setup wizard again on next launch.
Config File Location
The configuration file is stored at:
Windows
%LOCALAPPDATA%\Tag-AI\config.inimacOS
~/Library/Application Support/Tag-AI/config.iniLinux
~/.config/Tag-AI/config.iniWhile you can edit this file directly with a text editor, using the Configuration Editor is recommended to prevent syntax errors.
Example Configuration
[watch]
folder = C:/Users/username/Pictures, D:/Photos
[image_processing]
target_max_size_mb = 10
[tagger_gemini]
gemini_api_key = AIzaSyDaz8o3OAiKzO10H3Ild9umVxk5qyFNi8Y
paid_version = false
max_tags = 50
confidence_threshold = 0.7
max_image_size = 1024
[tagger_ollama]
ollama_endpoint = http://localhost:11434/api/generate
model_name = llava
[script_to_run]
llm_script = local
[license]
license_key = 6SP2-8CRS-TMTA-H21Y
[setup]
setup_complete = true
gpu_type = NVIDIA
gpu_acceleration = true
gpu_setup_skipped = false