Model & LoRA Library

Know Which Model Made Every Result

Stop guessing which LoRA or checkpoint produced your best work. Numonic auto-detects every model from your ComfyUI workflows and builds a browsable, searchable library — with full usage history.

  • Auto-Detection
  • Usage Tracking
  • Base Model Inference

How It Works

From import to browsable library in seconds

Auto-Detection

Import a ComfyUI image and Numonic extracts every model reference — LoRAs, checkpoints, VAEs, CLIP encoders, ControlNets. No manual tagging.

Browsable Library

All detected models appear in a filterable list organized by architecture and type. Search by name, filter by base model (SDXL, Flux, SD 1.5), or browse by type.

Usage Tracking

See every asset generated with a specific model. Click through to view the full gallery — same thumbnails, same navigation as your main asset library.

Comprehensive Model Detection

Numonic reads the embedded ComfyUI workflow JSON and identifies every model reference in the graph — across all node types.

Base model architecture is inferred automatically from naming patterns, covering 30+ CivitAI conventions. Names are normalized so dreamshaper_8.safetensors and DreamShaper v8 resolve to the same entry.

Every model type detected across your library appears once — deduplicated, labelled, and linked to every asset it touched.

LoRA
Checkpoint
UNET
VAE
ControlNet
CLIP
Upscaler
Embedding
IP-Adapter

For Creators and Teams

Built for how creative work actually happens

For Individual Creators

Track which LoRA gives you the best results. Record trigger words so you don't forget them. Link to CivitAI or HuggingFace source pages. Build a personal model reference library that grows automatically.

For Studios and Teams

Know what models your team is using across projects. Standardize on approved checkpoints. Track model versions and quality over time. Share model knowledge without tribal knowledge.

Start Building Your Model Library

Import your first ComfyUI image. Numonic does the rest.