AI & Models

Ollama Connection

Analyses photos, generates descriptions, and suggests filenames. Must support image input.
Enables natural-language search queries. Without this, search uses basic keyword matching.

Processing Modes


Advanced: Prompt Library Customise the prompts sent to AI models

Vision Prompt

The main prompt sent to the AI model when processing photos
Context Injection Template Injected into the vision prompt when you provide processing context

Rename & Filenames

Rename Mode

0.6 For auto-low-confidence mode: below this threshold triggers review

Filename Template

Placeholders: {date}, {location}, {description}, {camera}

Metadata Options

Scanning & Scheduling

Watch Mode

Uses OS-level notifications (inotify/FSEvents) for instant detection. Use manual Scan for a full directory walk.

Processing

Be careful with VRAM — more workers = more parallel model calls

Schedule

Only process during a daily time window
Scans still discover new files anytime.

Dashboard

Photo Showcase

Display two random photos at the bottom of the dashboard. Refreshes on each page load.
Leave empty to show any processed photo at random.
How often each showcase photo changes.
How many times faster tiles change on the mosaic vs the showcase.
How long each photo dissolves into the next.

Storage & Security

Thumbnails

Destructive Mode

Controls whether delete and trash operations are available. Disable to prevent accidental data loss.

Security

Database Management

Saves to backups folder and downloads a copy

Check database records against files on disk

Restore from a backup — current data will be replaced