1) What this app does
TomsBGRemover removes backgrounds from images using AI, entirely offline. It produces clean cut-outs with transparency (great for PNG exports), and includes dedicated modes for sticker creation and automatic “best settings” selection.
- Transparent PNG cut-outs
- Stickers (white outline, print-and-cut)
- Product photos for online stores
- Thumbnails and social media assets
- Processing large image folders (batch mode)
- No image uploads
- No cloud processing
- No account required
- No tracking
2) System requirements
- Windows 10 or Windows 11 (64-bit)
- 8 GB RAM recommended
- Modern CPU recommended for faster processing
- Uses Windows DirectML when available
- If not supported, the app automatically uses CPU
- Results are the same — GPU mainly improves speed
3) First-time setup
- Extract the ZIP
Right-click the ZIP → Extract All. Don’t run the app from inside the ZIP. - Keep the folder structure intact
The app expects certain folders (for example the AI model folder). Don’t rename or move internal folders. - Run the app
Double-click TomsBGRemover.exe.
4) Quick start (single image in ~60 seconds)
- Click Load Image and choose a JPG or PNG.
- Select a mode:
- Auto AI (recommended) – best general results
- Sticker Mode – creates a sticker-ready outline
- Normal AI – classic clean background removal
- Click Remove Background.
- Preview the result on the right.
- Click Save Image (usually exports a transparent PNG).
5) Single image workflow
Standard workflow
- Load your image.
- Choose a processing mode.
- Process the image.
- Save the output.
6) Batch workflow (folders)
Batch mode is for processing a whole folder of images automatically — great for product catalogs, sticker packs, and bulk photo prep.
- Click Select Input Folder and choose the folder containing your images.
- Click Select Output Folder and choose where results should be saved.
- Select your mode (Auto AI is best for mixed folders).
- Click Start Batch.
7) Modes explained
Best for photos and clean product shots
- Removes background with AI
- Conservative, clean edges
- Great default for most photos
Best for sticker creation and print-and-cut
- Removes background
- Cleans mask (removes specks/holes)
- Keeps the main subject
- Adds a smooth white outline
Auto AI analyzes each image and chooses the best combination of settings automatically. It can decide whether the image should be treated more like a sticker/graphic or a photo subject.
- Great for mixed folders
- Locks manual advanced settings (because the AI is making decisions)
- Designed for “set it and forget it” workflows
8) Advanced settings (optional)
Advanced settings let you tune quality and output behavior. These settings are saved and restored automatically the next time you open the app.
- Mask sensitivity – how aggressively the subject is detected
- Feather – edge softness (useful for hair/fur)
- Edge cleanup – reduces jagged/noisy edges
- Color spill suppression – helps reduce background color bleed on edges
- Transparent – standard cut-out result
- Solid color – choose a background color
- Image background – replace with an image
9) Output files
- Typically saves as PNG
- Transparency preserved
- Saved to your selected output folder
- Designed to safely overwrite files when needed
- If a file is in use elsewhere, Windows may block overwriting
10) GPU vs CPU
TomsBGRemover attempts to use GPU acceleration through Windows DirectML when available. If your GPU or drivers do not support it, the app automatically uses CPU mode.
- Faster processing
- Same output quality
- Automatically falls back to CPU
- No errors required
- Same output quality
11) Tips for best results
- Use high-resolution images when possible
- Ensure the subject is in focus
- Try a small amount of feathering for hair/fur
- Use edge cleanup if edges look jagged
- Use Sticker Mode for clean outlines
- Use Auto AI if you’re unsure
- If tiny bits are missing, increase sensitivity slightly
- If extra noise appears, enable cleanup
12) Known limitations (current version)
- AI Upscaling: Current upscale is basic resizing (true AI super-resolution is planned).
- Multi-subject separation: The app focuses on the main subject in many workflows (especially sticker cleanup).
- GPU support: Depends on Windows DirectML and your driver support — CPU fallback is normal on some systems.
13) Troubleshooting & support tickets
This section is designed so users can send us a clean, useful support request without back-and-forth. If you’re reporting an issue, please follow the checklist below.
- Close and reopen the app
- Re-extract the ZIP (corrupted extract causes weird issues)
- Make sure you are not running the app from inside the ZIP
- Confirm the app folders (like model folders) were not renamed or moved
- Try processing a different image to confirm if the issue is image-specific
Common issues
- Click More info → Run anyway
- Ensure the ZIP is extracted first
- If your antivirus quarantined files, restore them and add the app folder as an exception (if you trust the source)
- Enable Edge cleanup
- Increase Feather slightly for soft edges (hair/fur)
- Try Auto AI Mode for automatic tuning
- For cartoons/logos, use Sticker Mode
- Close any program that might have the output image open
- Try saving with a new filename
- Make sure the output folder is writable (not a protected system folder)
- Large images take longer — try smaller resolution output
- CPU mode is normal on some systems
- Try a smaller test batch (5–10 images) to confirm stability
How to submit a support ticket (copy/paste template)
Please copy this template and fill it out:
TomsBGRemover Support Request 1) App Version: (Shown in the About section) 2) Windows Version: (Windows 10/11 + edition if known) 3) CPU + RAM: (Example: Ryzen 5, 16GB) 4) GPU (if any): (Example: NVIDIA GTX 1650 / AMD RX xxxx / Intel) 5) What were you doing when the issue happened? (Single image or Batch? Which mode?) 6) What did you expect to happen? 7) What actually happened? (Error message? Crash? Wrong output? Slow?) 8) Steps to reproduce: 1. 2. 3. 9) Sample image(s): (Attach 1–2 images if possible) 10) Output result image(s): (Attach the output if relevant) 11) Any logs or screenshots: (Attach if available)