A lightweight, local, browser-based gallery that remembers
exactly how every image or video was generated.
SmartGallery turns your ComfyUI output folder into a fast, searchable, mobile-friendly web gallery.
Every generated file (image or video) is automatically linked to its exact ComfyUI workflow, even if:
- ComfyUI is not running
- the file was generated weeks or months ago
- filenames are meaningless
Everything runs fully offline, locally, with no cloud and no tracking.
Who is this for?
ComfyUI users who generate a lot, iterate fast, and want to never lose a workflow again.
If you use ComfyUI seriously, you probably faced this:
- Thousands of outputs with generic filenames
- Forgotten workflows
- Hard to search by what you actually remember
- No usable way to browse from your phone
SmartGallery turns your output folder into a living memory of your creative process.
| Desktop Interface |
|
| Fast, clean desktop gallery with workflow-aware search |
| Mobile Interface | Node Summary |
|
|
| Fully usable on mobile | Instant workflow recall with visual node summary |
- Search & Filter: Find files by keywords, specific models/LoRAs, file extension, date range, and more.
- Full Workflow Access: View node summary, copy to clipboard, or download JSON for any PNG, JPG, WebP, WebM or MP4.
- File Manager Operations: Select multiple files to delete, move, copy or re-scan in bulk. Add and rename folders.
- Mobile-First Experience Optimized UI for desktop, tablet, and smartphone.
- Compare Mode: Professional side-by-side comparison tool for images and videos with synchronized zoom, rotate and parameter diff.
- Video Storyboard Overview Analyze video content with a clean 11-frame grid covering the entire duration.
- External Folder Linking: Mount external hard drives or network paths directly into the gallery root, including media not generated by ComfyUI.
- Auto-Watch: Automatically refreshes the gallery when new files are detected.
- Cross-platform: Windows, Linux, macOS, and Docker support. Completely platform agnostic.
- Fully Offline: Works even when ComfyUI is not running.
Actively developed — frequent updates focused on real ComfyUI workflows.
In this release, along with several other features, we are introducing a brand-new tool for analyzing videos
The 11-frame Grid covering the entire duration from Start to the Last Frame
- Video Storyboard & Analysis Instantly analyze video content with a clean 11-frame Grid covering the entire duration from Start to the True Last Frame.
- Options Menu: New persistent
⚙️ Optionsmenu (Desktop/Mobile) to manage core gallery settings. - Thumbnail Grid Size: Added a new toggle in the Options menu (
⚙️) allowing users to switch between Normal and Compact view on desktop. - Focus Mode: A new streamlined view. Hides UI clutter and changes click behavior to "Select Only" for rapid batching. Accessible via the
⚡button orQkey. - Shortcuts Button: Added many shortcuts and a dedicated
? Shortcutsbutton in the desktop header for Quick Actions description. - Generation Dashboard: Added a high-fidelity summary panel at the top of the Node Summary to show Seed, Model, Steps, and Prompts at a glance.
- Smart Video Grid: Completely rewritten "Play/Pause on Scroll" engine. Videos strictly pause when leaving the viewport, drastically reducing CPU/GPU usage.
- Video Autoplay Control: A session-based toggle to enable/disable video autoplay in the grid. (Default: OFF).
Pkey shortcut to quickly toggle the Video Autoplay setting. - Dynamic UX for Videos:
- On Desktop, when Autoplay is OFF, a small ▶ icon appears in the corner. Clicking it plays the video in-grid for quick preview.
- On Mobile, the thumbnail is fully clickable to open the Lightbox (Click-to-Open).
- Auto-Consistency: The system forces a robust Full Sync on every startup to remove "ghost" files deleted externally via OS.
This is a curated summary — see the 👉 full CHANGELOG for all technical changes.
Select your platform
Each quick install shows only the relevant steps for that platform.
Windows (Python)
Option A: Using Git (Recommended)
git clone https://github.com/biagiomaf/smart-comfyui-gallery
cd smart-comfyui-galleryOption B: No Git (Manual Download)
Download the latest Source code (zip) from Releases, extract it, and open a terminal inside the folder.
Then, setup the environment:
python -m venv venv
venv\Scripts\activate
pip install -r requirements.txtCreate a new file named run_smartgallery.bat inside the folder and paste this content.
/ even on Windows.
@echo off
cd /d %~dp0
call venv\Scripts\activate.bat
:: --- CONFIGURATION ---
:: REPLACE these paths with your actual folders.
:: NOTE: Use forward slashes (/) for paths (e.g., C:/ComfyUI/output)
set "BASE_OUTPUT_PATH=C:/Path/To/ComfyUI/output"
set "BASE_INPUT_PATH=C:/Path/To/ComfyUI/input"
set "BASE_SMARTGALLERY_PATH=C:/Path/To/ComfyUI/output"
:: If ffmpeg is not in your system PATH, point to ffprobe.exe here:
set "FFPROBE_MANUAL_PATH=C:/Path/To/ffmpeg/bin/ffprobe.exe"
set SERVER_PORT=8189
:: --- START ---
python smartgallery.py
pauseDouble-click run_smartgallery.bat to start.
If you installed via Git:
cd smart-comfyui-gallery
git pull
venv\Scripts\activate
pip install -r requirements.txtIf you downloaded the ZIP: Download the new version, extract it, and copy your run_smartgallery.bat into the new folder.
macOS (Python)
Option A: Using Git (Recommended)
git clone https://github.com/biagiomaf/smart-comfyui-gallery
cd smart-comfyui-galleryOption B: No Git (Manual Download)
Download the latest Source code (tar.gz) from Releases, extract it, and open a terminal inside the folder.
Then, setup the environment:
python3 -m venv venv
source venv/bin/activate
pip install -r requirements.txtCreate a file named run_smartgallery.sh, make it executable (chmod +x run_smartgallery.sh), and paste this content:
#!/bin/bash
source venv/bin/activate
# Increase open files limit (fix for "Too many open files" on macOS)
ulimit -n 4096
# --- CONFIGURATION ---
# REPLACE these paths with your actual folders.
export BASE_OUTPUT_PATH="$HOME/ComfyUI/output"
export BASE_INPUT_PATH="$HOME/ComfyUI/input"
export BASE_SMARTGALLERY_PATH="$HOME/ComfyUI/output"
# Ensure ffprobe is installed (brew install ffmpeg)
export FFPROBE_MANUAL_PATH="/usr/bin/ffprobe"
export SERVER_PORT=8189
# --- START ---
python smartgallery.pyRun it with: ./run_smartgallery.sh
cd smart-comfyui-gallery
git pull
source venv/bin/activate
pip install -r requirements.txtIf you downloaded the tar.gz: Download the new version, extract it, and copy your run_smartgallery.sh into the new folder.
Linux (Python)
Option A: Using Git (Recommended)
git clone https://github.com/biagiomaf/smart-comfyui-gallery
cd smart-comfyui-galleryOption B: No Git
Download Source code from Releases.
Then, setup the environment:
python3 -m venv venv
source venv/bin/activate
pip install -r requirements.txtCreate a file named run_smartgallery.sh, make it executable (chmod +x run_smartgallery.sh), and paste this content:
#!/bin/bash
source venv/bin/activate
# --- CONFIGURATION ---
# REPLACE these paths with your actual folders.
export BASE_OUTPUT_PATH="$HOME/ComfyUI/output"
export BASE_INPUT_PATH="$HOME/ComfyUI/input"
export BASE_SMARTGALLERY_PATH="$HOME/ComfyUI/output"
export FFPROBE_MANUAL_PATH="/usr/bin/ffprobe"
export SERVER_PORT=8189
# --- START ---
python smartgallery.pyRun it with: ./run_smartgallery.sh
cd smart-comfyui-gallery
git pull
source venv/bin/activate
pip install -r requirements.txtIf you downloaded the source code manually: Download the new version, extract it, and copy your run_smartgallery.sh into the new folder.
Docker
Replace the paths on the left side of the : with your actual host paths.
docker run \
--name smartgallery \
-v /your/host/output:/mnt/output \
-v /your/host/input:/mnt/input \
-v /your/host/SmartGallery:/mnt/SmartGallery \
-e BASE_OUTPUT_PATH=/mnt/output \
-e BASE_INPUT_PATH=/mnt/input \
-e BASE_SMARTGALLERY_PATH=/mnt/SmartGallery \
-p 8189:8189 \
-e WANTED_UID=`id -u` \
-e WANTED_GID=`id -g` \
mmartial/smart-comfyui-gallery# 1. Pull the latest image
docker pull mmartial/smart-comfyui-gallery
# 2. Stop and remove the old container
docker stop smartgallery && docker rm smartgallery
# 3. Run the 'docker run' command again (see above)Want to run SmartGallery in a containerized environment? We've got you covered!
🎖️ Special Thanks: A huge shout-out to Martial Michel for orchestrating the Docker support and contributing to the core application logic.
Docker deployment provides isolation, easier deployment, and consistent environments across different systems. However, it requires some familiarity with Docker concepts.
🗄️ Pre-built images
Pre-built images are available on DockerHub at mmartial/smart-comfyui-gallery and Unraid's Community Apps.
Full Docker guide: 👉 docs/DOCKER_HELP.md
http://127.0.0.1:8189/galleryout
For advanced configuration: 👉 Complete Installation Guide (Docker Compose, reverse proxy, ffmpeg, scripts, and more)
Testing cutting-edge features before official releases?
The /experiments folder contains beta versions and hotfixes under active development.
SmartGallery is designed to stay lightweight by default.
Advanced AI-powered features (semantic search, natural language queries) will be provided by a separate optional service:
- Fully optional
- Runs locally
- Separate Docker container or Python environment
- No impact on the core gallery if not installed
The AI service is currently under development and not released yet.
- Local-first
- Privacy-first
- Minimal dependencies
- No forced upgrades
- No vendor lock-in
Issues, ideas, and pull requests are welcome.
- Open an issue
- Fork → branch → PR
If SmartGallery improves your ComfyUI workflow, consider giving the repository a ⭐.
It helps visibility, motivates future development, and costs nothing.
MIT License — see LICENSE
Made for the ComfyUI community
