🧩 ComfyDash – Local Model Catalog Dashboard

ComfyDash helps you make sense of all your downloaded checkpoints, LoRAs, and embeddings by scanning your ComfyUI directories and presenting them in a clear, searchable dashboard. It's a lightweight local catalog and management tool for ComfyUI model data.

Do you know the issue? You download checkpoints, LoRAs, and embeddings, use them for a while, and later wonder what they actually do. ComfyDash was created to solve exactly that — giving you a clear overview of everything stored in your local ComfyUI setup.

ComfyDash combines a ⚙️ Python‑based scanner that generates a structured catalog.json file with a modern 💻 React frontend built using Vite and TailwindCSS.

---

🚀 Features

🧩 Scanner (Python CLI)

• 🔍 Scans any ComfyUI model directory (Checkpoints, LoRAs, Embeddings)
• 🧾 Generates a unified catalog.json file
• 📏 Captures file size, modification date, type, and stable ID
• 🚫 Automatically skips missing or invalid files
• 🔐 Extracts metadata from Safetensors files (triggers, tags, base model, CivitAI URLs)
• 🎯 NEW in v1.3: Improved architecture detection for SDXL, Pony, Illustrious, and Cascade models

🖥️ Dashboard (React + Tailwind)

• 🗂️ Accordions per model type (Checkpoint / LoRA / Embedding) with "Select All" functionality
• 🔎 Search & filter (client‑side, no backend required)
• ✏️ Inline editing for model name, CivitAI link, trigger tags, base model, etc.
• 💾 Local persistence of all edits via localStorage
• 🧠 Heuristic suitability detection (📷 Realistic / ✏️ Drawing) with manual override
• ⭐ Favorites system
• 📌 Sticky header for improved navigation
• 🤖 Automatic metadata extraction from Safetensors (trigger words, tags, base model)
• 🌐 CivitAI integration - search selected models on CivitAI and auto-fill metadata
• 📊 NEW in v2.0.1: Visual progress overlay during CivitAI searches with file-by-file status
• 📏 NEW in v2.0.1: Resizable columns - drag column borders to adjust widths, settings saved per model type
• 🎨 NEW in v1.3: Optimized column layout - wider model names, compact file paths
• 🖱️ NEW in v1.3: Click column headers to sort - visual indicators show sort direction
• 🏗️ NEW in v1.3: Pony/Illustrious workflow template included
• 🚀 NEW in v2.0: Launch ComfyUI directly from the dashboard with one click
• 🐍 NEW in v2.0: Conda environment support for ComfyUI startup
• ✅ NEW in v2.0: Real-time ComfyUI status detection and smart button switching

---

🧰 Installation

🧾 Requirements

• 🧩 Node.js ≥ 18.x (includes npm)
• 🐍 Python 3.12+

🪟 Quick Start (Windows)

ComfyDash can now be launched with a single double‑click.

1️⃣ Clone and setup

git clone https://github.com/max-dingsda/ComfyDash
cd ComfyDash
cd comfydash
npm install

2️⃣ Start ComfyDash

Run the PowerShell script in the project root:

start_comfydash.ps1

This will:

• 🧠 start the local Python API (mini_server.py)
• 🧰 start the Vite dev server (npm run dev)
• 🌐 open your browser at http://localhost:5173

💡 Alternative: run start_comfydash.bat for an even quicker double‑click start.

---

🧪 Usage

Basic Workflow

1. 🌍 Open the dashboard in your browser
2. 🔍 Click "Detect API" to find the running backend
3. 📂 Enter your ComfyUI root path (e.g., F:\AI\ComfyUI)
4. ⚡ Click "Scan now" to scan your models
5. 🔎 Browse, search, and organize your models
6. 💾 All edits are automatically saved in your browser

CivitAI Integration (v1.2)

1. ✅ Select models using the checkboxes
2. 🔍 Click "Find selected on CivitAI"
3. ⏳ Wait for the search to complete (rate-limited to ~2 models/second)
4. 🎉 Metadata (title, URL, triggers) is automatically filled in

Note: CivitAI search uses file hashes and may take several minutes for many models. Manual data will be overwritten by CivitAI data.

ComfyUI Launch (v2.0)

1. 📂 Enter your ComfyUI root path
2. 🐍 (Optional) Specify a Conda environment name (e.g., "comfyui")
3. 🚀 Click "Open ComfyUI" to start ComfyUI
4. ⏳ Wait ~10-30 seconds for startup
5. ✅ Green button appears when ready - click to open ComfyUI in new tab

Note: If you use a Conda environment for ComfyUI, enter the environment name. Leave empty to use system Python. ComfyUI will launch in a visible console window so you can monitor startup and errors.

🐍 Running the scanner manually

python scanner/main.py --root "<path_to_your_ComfyUI_folder>" --output "catalog.json"

🌐 Mini API (for integration)

The local API mini_server.py allows the dashboard to run the scanner directly:

python mini_server.py --host 127.0.0.1 --port 8000

Endpoints

• 🩺 GET /health → { ok: true }
• 🧭 POST /scan → { root: "F:\\AI\\ComfyUI", output: "optional\\catalog.json" }
• 🌐 POST /enrich-civitai → { path: "path/to/model.safetensors" }
• 🔍 GET /comfyui/status → { ok: true, running: true/false }
• 🚀 POST /comfyui/start → { root: "F:\\AI\\ComfyUI", port: 8188, conda_env: "optional" }

---

🧭 Roadmap

Version	Focus	Enhancements
1.0 ✅	MVP Release	Stable local version with all core functionality
1.1 ✅	Usability	Run scanner from UI, API auto-detect, local annotations, base-model badges (SD 1.5 / SDXL / FLUX / PONY)
1.2 ✅	Metadata	Extract Safetensors metadata, CivitAI API integration, manual override for suitability flags, sticky header, "Select All"
1.3 ✅	UX & Polish	Improved architecture detection, sortable columns, optimized layout, workflow templates (prep for v2.0)
2.0 ✅	Workflows	Launch ComfyUI directly from dashboard, conda environment support, real-time status detection
2.0.1 ✅	Bug Fixes & UX	Fixed CivitAI API integration (missing /enrich-civitai endpoint), added progress overlay for CivitAI searches, resizable columns with per-type persistence
2.0.2 ✅	Bug Fixes	Fixed input field focus loss during typing, fixed sorting to check manual annotations and use case-insensitive comparison
2.1	Workflows	🔜 Launch ComfyUI with preconfigured workflow templates

---

🤝 Contributing

Contributions are welcome! Feel free to open issues or submit pull requests.

---

📜 License

MIT License - see LICENSE file for details.

---

🙏 Acknowledgments

Built with ❤️ for the ComfyUI community.