Configuration
Customize Smart Asset Finder settings for optimal performance and workflow.
Access Settings
Go to Edit → Editor Preferences → Plugins → Smart Asset Finder
Settings Overview
Scanning Settings
Scan Paths
Default: ["/Game"]
Directories to scan for assets. Multiple paths supported.
Examples:
/Game # Scan entire Game folder
/Game/Characters # Scan only Characters
/MyPlugin/Content # Scan plugin content
Use case: Limit scanning to specific folders to speed up index builds.
Target Asset Classes
Default: All supported types
Select which asset types to include in the index.
Supported types:
- Texture2D
- StaticMesh
- SkeletalMesh
- Material
- MaterialInstance
- ParticleSystem
- PaperSprite
Use case: Exclude asset types you don’t need to reduce index size.
Exclude Paths
Default: Empty
Paths to exclude from scanning.
Example:
/Game/Dev # Exclude development assets
/Game/Test # Exclude test assets
/Game/Deprecated # Exclude old assets
Include Engine Content
Default: false
Include engine content (/Engine folder) in scans.
Warning: Significantly increases index size and build time.
Use case: Only enable if you need to search engine assets.
Performance Settings
Rebuild Batch Size
Default: 10
Range: 1-100
Number of assets to process per batch during manual rebuild.
Guidelines:
- Lower (1-5): Less memory, slower speed
- Medium (10-20): Balanced (recommended)
- Higher (50-100): Faster, more memory
When to adjust:
- Low memory: Reduce to 5
- High-end CPU: Increase to 20-50
- Very large projects: Reduce to prevent memory issues
ONNX Inference Threads
Default: 0 (auto-detect)
Range: 0-16
Number of CPU threads for AI inference.
Options:
- 0 (Auto): Uses 50% of CPU cores (recommended)
- 1 (Single): Low-end CPUs or background processing
- 2-8: Fixed thread count for consistent performance
- 8-16: High-end workstations
Guidelines by CPU:
| CPU Cores | Recommended Setting |
|---|---|
| 4 cores | 0 (auto) or 2 |
| 6-8 cores | 0 (auto) or 4 |
| 8-12 cores | 0 (auto) or 6 |
| 12+ cores | 0 (auto) or 8 |
Enable Detailed Logging
Default: false
Enable verbose logging for debugging.
Warning: May impact performance. Use only for troubleshooting.
Output: Check Output Log for detailed plugin messages.
Auto-Update Settings
Enable Auto-Update
Default: true
Automatically monitor and update index when assets change.
Recommended: Keep enabled for active development.
Auto Update Batch Size
Default: 5
Range: 1-50
Assets processed per batch during auto-update.
Guidelines:
- Lower (1-5): More responsive, frequent updates
- Higher (10-20): Fewer interruptions, slower to reflect changes
Auto Update Debounce Time
Default: 3.0 seconds
Range: 0.5-60.0 seconds
Delay after last asset change before processing.
Purpose: Prevents continuous re-indexing during rapid editing (e.g., bulk import).
Guidelines:
- Short (1-2s): See changes quickly
- Medium (3-5s): Balanced (recommended)
- Long (10s+): Reduce CPU usage during heavy editing
Max Queue Size
Default: 100
Range: 10-1000
Maximum assets to queue before warning user.
Purpose: Prevents memory issues during bulk imports.
When to adjust:
- Frequent bulk imports: Increase to 500
- Low memory: Keep default or reduce to 50
Pause During PIE
Default: true
Pause auto-update during Play-In-Editor sessions.
Recommended: Keep enabled to avoid performance impact during testing.
Storage Settings
Index Storage Path
Default: "SmartAssetFinder" (relative to Saved/)
Directory for index storage.
Full path: YourProject/Saved/SmartAssetFinder/
Use case: Change if you have custom backup/sync workflows.
Configuration Profiles
Recommended Profiles
Default (Balanced)
RebuildBatchSize: 10
OnnxInferenceThreads: 0 (auto)
AutoUpdateBatchSize: 5
AutoUpdateDebounceTime: 3.0
EnableAutoUpdate: true
PauseDuringPIE: true
High Performance (Fast indexing)
RebuildBatchSize: 50
OnnxInferenceThreads: 8
AutoUpdateBatchSize: 10
AutoUpdateDebounceTime: 1.0
EnableAutoUpdate: true
PauseDuringPIE: true
Use when: High-end workstation, want fastest index builds.
Low Memory (Resource-constrained)
RebuildBatchSize: 5
OnnxInferenceThreads: 1
AutoUpdateBatchSize: 3
AutoUpdateDebounceTime: 5.0
EnableAutoUpdate: true
PauseDuringPIE: true
Use when: 4GB RAM, low-end CPU, or background processing.
Large Projects (10k+ assets)
RebuildBatchSize: 10
OnnxInferenceThreads: 0 (auto)
AutoUpdateBatchSize: 5
AutoUpdateDebounceTime: 10.0
EnableAutoUpdate: false # Manually trigger updates
PauseDuringPIE: true
Use when: Very large projects where auto-update is impractical.
Optimizing for Your Hardware
High-End Workstation
- CPU: 12+ cores
- RAM: 32GB+
- Settings: High Performance profile
- Expected speed: ~2 seconds per 100 assets
Standard Development PC
- CPU: 6-8 cores
- RAM: 16GB
- Settings: Default profile
- Expected speed: ~1 second per 50 assets
Low-End / Laptop
- CPU: 4 cores
- RAM: 8GB
- Settings: Low Memory profile
- Expected speed: ~2-3 seconds per 50 assets
Troubleshooting Settings
Index builds are too slow
Solutions:
- Increase
RebuildBatchSizeto 20-50 - Increase
OnnxInferenceThreadsto 4-8 - Limit
ScanPathsto necessary folders
Editor freezes during index build
Solutions:
- Reduce
RebuildBatchSizeto 5 - Reduce
OnnxInferenceThreadsto 1-2 - Close other applications
Auto-update is too aggressive
Solutions:
- Increase
AutoUpdateDebounceTimeto 10-30 seconds - Reduce
AutoUpdateBatchSizeto 3 - Temporarily disable auto-update during heavy editing
Out of memory errors
Solutions:
- Reduce
RebuildBatchSizeto 3-5 - Reduce
AutoUpdateBatchSizeto 3 - Exclude large asset folders via
ExcludePaths - Disable
IncludeEngineContent
Settings not taking effect
Solutions:
- Restart the editor after changing settings
- Delete index folder (
Saved/SmartAssetFinder/) and rebuild - Check Output Log for errors
Best Practices
- Start with defaults: Only adjust if you have specific needs
- Test performance: Try different batch sizes to find optimal for your hardware
- Monitor memory: Watch Task Manager during index builds
- Use exclusions: Exclude temp/test folders to reduce index size
- Enable auto-update: Keep index synchronized during development
- Disable during imports: Turn off auto-update before bulk imports, re-enable after