[Script request]: SnapOtter 'image toolkit'

[easaw]

Application Name

SnapOtter

Website

https://github.com/snapotter-hq/snapotter

Description

I don't want to have to upload images to 'the web' also it takes time, and I don't really want to open apps like Photoshop or Affinity to do simple tasks, especially in bulk (I'm also not paying for Photoshop. Eat sh*t Adobe...).

However, I should note; I am not interested in the AI tools, that is not why I am requesting this app. It's purely about quick editing.

From GitHub:

"🦦The open-source image toolkit that makes paid tools nervous. Every image tool you need. Your images stay yours."

• 50+ image tools: Resize, crop, compress, convert, watermark, color adjust, beautify screenshots, generate memes, vectorize, create GIFs, find duplicates, generate passport photos, and more. Supports 55+ input formats (including 23 camera RAW formats) and 14 output formats
• Image editor: Layer-based editor with brushes, shapes, adjustments, filters, curves, and keyboard shortcuts. Runs in your browser, processes on your hardware
• Local AI: Remove backgrounds, upscale images, restore and colorize old photos, erase objects, blur faces, enhance faces, extract text (OCR), expand canvas, fix transparency. All on your hardware, no internet required
• OIDC / SSO: Login with Google, GitHub, Okta, or any OpenID Connect provider
• 20 languages: Arabic, Chinese, Czech, Dutch, French, German, Hindi, Indonesian, Italian, Japanese, Korean, Polish, Portuguese, Russian, Spanish, Thai, Turkish, Ukrainian, Vietnamese. RTL support for Arabic
• Pipelines: Chain tools into reusable workflows with unlimited steps. Import/export as JSON. Batch process unlimited images at once
• REST API: Every tool available via API with API key auth. Interactive docs at /api/docs
• Single container: One docker run, no Redis, no Postgres, no external services
• Multi-arch: Runs on AMD64 and ARM64 (Intel, Apple Silicon, Raspberry Pi)
• Privacy first: Your images never leave your network. SnapOtter asks once whether you'd like to share anonymous product analytics (which tools are used, errors encountered, never file data). Change anytime in Settings, or set ANALYTICS_ENABLED=false to disable completely

Also: https://snapotter.com/

Thank you.

Due Diligence

(searched existing)

• I have searched existing scripts and found no duplicates.
• I have searched existing discussions and found no duplicate requests.
• The application requested has 600+ stars on Github (if applicable), is older than 6 months, actively maintained and has release tarballs published.
• I understand that not all applications will be accepted due to various reasons and criteria by the community-scripts ORG.

[dhoitink]

Yes Please, was about to ask for this!

[Aloe-recite]

Also looking for this, although I'm not sure that it fits this community's requirements: 1.5k stars on GH, which is remarkable, but the project is still very new (2 months old).

[MickLesk]

community-scripts/ProxmoxVED#1867

[MickLesk]

can you spin up the lxc again? (New Install)

[Aloe-recite]

now it works and installs as requested! still an issue though: if you install then uninstall a plugin, it says "uninstalled" but used space does not decrease. if you install it again, it completes immediately. seems like "uninstall" does not actually delete the files.

[michelroegl-brunner]

now it works and installs as requested! still an issue though: if you install then uninstall a plugin, it says "uninstalled" but used space does not decrease. if you install it again, it completes immediately. seems like "uninstall" does not actually delete the files.

That is a issue for the devs of the app then.

[Aloe-recite]

alright, it still works though and just in case files can be deleted manually I assume?

[MickLesk]

they are stored in /opt/snappotter_data

[Aloe-recite]

@easaw, @dhoitink can you please test it too and give feedback?

[dhoitink]

Not sure of this is Snapotter or script issue; When trying to install AI tools

Remove background: Failed to download 8 model(s): rembg-u2net, rembg-bria-rmbg, rembg-birefnet-general-lite, rembg-birefnet-portrait, rembg-isnet-general-use, rembg-birefnet-general, rembg-birefnet-hr-matting, rembg-birefnet-matting

or when trying to download the AI Image upscaling

pip install failed for 'realesrgan==0.3.0': error: subprocess-exited-with-error × Getting requirements to build wheel did not run successfully. │ exit code: 1 ╰─> [24 lines of output] Traceback (most recent call last): File "/opt/snapotter_data/ai/venv/lib/python3.13/site-packages/pip/_vendor/pyproject_hooks/_in_process/_in_process.py", line 389, in main() ~~~~^^ File "/opt/snapotter_data/ai/venv/lib/python3.13/site-packages/pip/_vendor/pyproject_hooks/_in_process/_in_process.py", line 373, in main json_out["return_val"] = hook(**hook_input["kwargs"]) ~~~~^^^^^^^^^^^^^^^^^^^^^^^^ File "/opt/snapotter_data/ai/venv/lib/python3.13/site-packages/pip/_vendor/pyproject_hooks/_in_process/_in_process.py", line 143, in get_requires_for_build_wheel return hook(config_settings) File "/tmp/pip-build-env-au1_ffh6/overlay/lib/python3.13/site-packages/setuptools/build_meta.py", line 333, in get_requires_for_build_wheel return self._get_build_requires(config_settings, requirements=[]) ~~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/tmp/pip-build-env-au1_ffh6/overlay/lib/python3.13/site-packages/setuptools/build_meta.py", line 301, in _get_build_requires self.run_setup() ~~~~~~~~~~~~~~^^ File "/tmp/pip-build-env-au1_ffh6/overlay/lib/python3.13/site-packages/setuptools/build_meta.py", line 520, in run_setup super().run_setup(setup_script=setup_script) ~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/tmp/pip-build-env-au1_ffh6/overlay/lib/python3.13/site-packages/setuptools/build_meta.py", line 317, in run_setup exec(code, locals()) ~~~~^^^^^^^^^^^^^^^^ File "", line 149, in File "", line 79, in get_version KeyError: 'version' [end of output] note: This error originates from a subprocess, and is likely not a problem with pip. ERROR: Failed to build 'basicsr' when getting requirements to build wheel

💡 PVE Version 9.1.9 (Kernel: 7.0.2-2-pve)
🖥️ Operating System: debian
🌟 Version: 13
📦 Container Type: Unprivileged
🆔 Container ID: 113
🏠 Hostname: snapotter
💾 Disk Size: 30 GB
🧠 CPU Cores: 4
🛠️ RAM Size: 4096 MiB
🌉 Bridge: vmbr0
🗂️ FUSE Support: no
📦 Nesting: Enabled
🎮 GPU Passthrough: no

[snapotter-hq]

Hey everyone, SnapOtter maintainer here. Thanks for the reports and for testing the Proxmox script!

I've pushed fixes for the issues reported in this thread:

Uninstall not freeing disk space -- this was a bug in our uninstall logic. It only knew how to clean up models stored as direct file paths, but background removal (rembg) and OCR (PaddleOCR) models use a different download mechanism and were being skipped entirely. That's also why reinstalling completed instantly -- the files were never actually removed. Fixed now; uninstall properly deletes all model types.

realesrgan/basicsr pip install failure (KeyError: '__version__') -- the basicsr dependency fails to build under modern pip's build isolation. We now pre-install basicsr with --no-build-isolation before realesrgan, so pip finds it already satisfied and never hits the broken build path. This should resolve the upscale/enhance and photo restoration install failures.

AI model download failures (rembg) -- this is typically network-related (timeouts or rate limiting when pulling ONNX models from GitHub/HuggingFace). We've improved the error messages to surface the actual cause and suggest retrying. If you're in a resource-constrained LXC container, make sure you have stable connectivity and enough memory (4 GB minimum, 6 GB recommended for the larger bundles).

These changes will be in the next release. Appreciate the interest in getting SnapOtter into the community scripts -- happy to help with any other issues that come up.

[MickLesk]

Nice. Thx. Are you thinking about to Release prebuild archives? Like jotty (https://github.com/fccview/jotty ) or other Tools?

If you see our Script (sorry im just on mobile atm), we need to dirty fix following:

ln -sf /opt/snapotter /app
sed -i 's/mediapipe==0.10.21/mediapipe>=0.10.21/' /opt/snapotter/docker/feature-manifest.json

and the whole Archive is an easy pnpm build, so an prebuild should be easy. If you Interessted on this, ping us. We can Support.

And the next Part is the uv part. An uv.lock should be easy, so we can just run uv sync

Here is the current Installer:
https://github.com/community-scripts/ProxmoxVED/blob/main/install/snapotter-install.sh

[snapotter-hq]

Thanks for the suggestions and for sharing the install script!

Pre-built archives: Yes, we're going to start publishing these with releases. Makes sense to avoid the pnpm build step on the host. We'll take a look at how jotty does it. Happy to coordinate on the format that works best for the community scripts. I have added this in my to-do will try to incorporate this as early as the next release itself.

mediapipe version pin: Already fixed on our end. We've loosened all mediapipe pins from ==0.10.21 to >=0.10.21, so you'll be able to drop that sed from the install script once the next release is out.

ln -sf /opt/snapotter /app: Noted, we should make the base path configurable instead of hardcoding /app. Will look into this.

uv.lock: Worth exploring. Our AI feature installer currently uses pip internally, but supporting uv as the package manager is on our radar in the future. This may take time.

Pls keep a lookout for the next release and do let me know if anything else bugs you or if any of the issues persists.

Thanks for the support!

[MickLesk]

you can just ping @fccview - I think he's taken a step back for personal reasons, but I think he might be able to help, since he created the prebuilds specifically for us. The process is actually pretty "simple" and you can find it under "Workflows."

Thanks for letting me know about the other issues. I'm waiting for the next release, and then I'll just update the two scripts. :-)

[fccview]

Hello!

First of all snapotter looks awesome and i'm just about to give it a go, never heard of it but I'm always on the lookout for free photo edit tools.

I'm in a much better place now @MickLesk , things are going very well and I have more capacity to work on my apps so very happy to help in case is needed @snapotter-hq ❤️

This is the Jotty workflow for the tarball release https://github.com/fccview/jotty/blob/main/.github/workflows/prebuild-release.yml

Granted, it's fairly different stack, but at least you can take bits and bobs for the github side of things :)

[snapotter-hq]

Thanks @fccview, glad you like the project! ❤️ We actually already had prebuilt archives set up in our release pipeline (multi-arch, amd64 + arm64, with SHA256 checksums). Took a look at the jotty workflow for reference, ours follows a similar pattern. They'll be attached to the next release.

Appreciate the offer to help, will reach out if we hit any snags!

[dhoitink]

Hi snapotter-hq, Thanks for taking the time!

I just did a fresh (re)install of the script (via bash -c "$(curl -fsSL https://github.com/community-scripts/ProxmoxVED/raw/main/ct/snapotter.sh)"

Was this the correct path for the script? I'm getting the same error when trying to install AI Photo Restoration.

PVE Version 9.1.9 (Kernel: 7.0.2-2-pve)
🖥️ Operating System: debian
🌟 Version: 13
📦 Container Type: Unprivileged
💾 Disk Size: 50 GB
🧠 CPU Cores: 4
🛠️ RAM Size: 6000 MiB

pip install failed for 'realesrgan==0.3.0': error: subprocess-exited-with-error × Getting requirements to build wheel did not run successfully. │ exit code: 1 ╰─> [24 lines of output] Traceback (most recent call last): File "/opt/snapotter_data/ai/venv/lib/python3.13/site-packages/pip/_vendor/pyproject_hooks/_in_process/_in_process.py", line 389, in main() ~~~~^^ File "/opt/snapotter_data/ai/venv/lib/python3.13/site-packages/pip/_vendor/pyproject_hooks/_in_process/_in_process.py", line 373, in main json_out["return_val"] = hook(**hook_input["kwargs"]) ~~~~^^^^^^^^^^^^^^^^^^^^^^^^ File "/opt/snapotter_data/ai/venv/lib/python3.13/site-packages/pip/_vendor/pyproject_hooks/_in_process/_in_process.py", line 143, in get_requires_for_build_wheel return hook(config_settings) File "/tmp/pip-build-env-8chseqil/overlay/lib/python3.13/site-packages/setuptools/build_meta.py", line 333, in get_requires_for_build_wheel return self._get_build_requires(config_settings, requirements=[]) ~~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/tmp/pip-build-env-8chseqil/overlay/lib/python3.13/site-packages/setuptools/build_meta.py", line 301, in _get_build_requires self.run_setup() ~~~~~~~~~~~~~~^^ File "/tmp/pip-build-env-8chseqil/overlay/lib/python3.13/site-packages/setuptools/build_meta.py", line 520, in run_setup super().run_setup(setup_script=setup_script) ~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/tmp/pip-build-env-8chseqil/overlay/lib/python3.13/site-packages/setuptools/build_meta.py", line 317, in run_setup exec(code, locals()) ~~~~^^^^^^^^^^^^^^^^ File "", line 149, in File "", line 79, in get_version KeyError: 'version' [end of output] note: This error originates from a subprocess, and is likely not a problem with pip. ERROR: Failed to build 'basicsr' when getting requirements to build wheel

[snapotter-hq]

Thanks for testing again! The fix hadn't been released yet when you reinstalled, so you were still running the old code. Apologies for the confusion.

We've also since found the deeper root cause of your specific error. Looking at your traceback, the venv is using Python 3.13 (python3.13 in the path), and basicsr's setup.py is fundamentally incompatible with Python 3.13 due to a changed locals() behavior (PEP 667). On top of that, newer versions of setuptools also break it. There are no pre-built wheels for basicsr on PyPI, so every install hits the broken source build.

The fix we're shipping in the next release: we pin setuptools<75 and pre-install basicsr with --no-build-isolation before realesrgan, which gives it a compatible build environment regardless of the Python version. We've also loosened the mediapipe version pins so they resolve properly on Python 3.13.

Will try to post here once the release is out. Please keep a lookout for the new release should be one this weekend. If you can test again at that point, that would be really helpful!

[dhoitink]

Hi @snapotter-hq, will do! Just send a message when the release is out.

[snapotter-hq]

v1.17.2 is out with the prebuilt archives attached. @MickLesk @dhoitink @Aloe-recite

Grab them from the release page:

• snapotter-v1.17.2-linux-amd64.tar.gz + .sha256
• snapotter-v1.17.2-linux-arm64.tar.gz + .sha256

Frontend already built, deps pruned to production-only, so you should be able to skip the pnpm build entirely in the install script.

All the fixes from this thread are in too:

• Uninstall actually deletes all model types now (rembg, PaddleOCR were being skipped before)
• basicsr/realesrgan installs fixed (pre-installed with --no-build-isolation, works on Python 3.13)
• mediapipe pins loosened to >=0.10.21 -- @MickLesk you can drop the sed
• Better error messages when rembg model downloads fail

We're also attaching SBOM and Trivy scan files to every release now (sbom.cdx.json, sbom.spdx.json, trivy.json). Not something most people need, but they're there if you want to audit deps. There's a hardening guide on the docs site too.

Full release notes: https://github.com/snapotter-hq/SnapOtter/releases/tag/v1.17.2

@MickLesk if you get a chance to test the prebuilt archive flow, let me know if the install script needs anything else on our end.

[MickLesk]

I had a look at the archive and its a bit different from how Jotty does it. Jotty ships a dedicated runtime bundle (standalone build), so our installer can simply extract it and start it directly with Node. In SnapOtters (your) case, the archive still contains most of the monorepo structure plus production node_modules, so it's more of a production-ready source bundle than a standalone runtime package.

Thats not a blocker for us, but it means we still need to inspect the startup flow and adjust our installer/service accordingly. The good news is that we can already drop the pnpm install and build steps, which was the main goal. I'm currently testing the new release and will report back if we find anything else that could make the prebuilt flow even cleaner for downstream packagers like us. Thanks again for the fast turnaround.

[MickLesk]

Maybe it's a minor issue—I've adjusted the script accordingly—but it seems there's still a hard-coded path?

SnapOtter looks here:
/app/packages/ai/python/install_feature.py

But our native installation is located here:
/opt/snapotter/packages/ai/python/install_feature.py

"/opt/snapotter_data/ai/venv/bin/python3: can't open file '/app/packages/ai/python/install_feature.py': [Errno 2] No such file or directory"

[MickLesk]

and another issue, i thought the mediapipe issue is solved?

root@snapotter:/opt# uv pip install \
venv/bi>     --python /opt/snapotter_data/ai/venv/bin/python \
t>     -r /opt/snapotter/packages/ai/python/requirements.txt
Using Python 3.13.5 environment at: snapotter_data/ai/venv
  x No solution found when resolving dependencies:
  `-> Because mediapipe==0.10.21 has no wheels with a matching Python ABI tag (e.g., `cp313`) and you require mediapipe==0.10.21, we can conclude that your requirements are unsatisfiable.

[snapotter-hq]

The /app path issue: install script path and manifest path were both hardcoded. Path resolution is now relative to where the source files actually live, so native installs at /opt/snapotter work without the symlink workaround. You can drop the ln -sf from your script after the next release.

The mediapipe pin: the >=0.10.21 change made it into the feature manifest but not into requirements.txt or requirements-gpu.txt. Fixed in both now. 0.10.35 ships py3-none wheels so Python 3.13 resolves fine.

Patch release coming shortly. Will try to finish up testing after work today.

If you get a chance to test it too please do. :)

[dhoitink]

Is the Proxmox script also updated ?

[dhoitink]

I asked because I got the same error when installing AI / Photo Restoration after a new/fresh installation about an hour ago from

bash -c "$(curl -fsSL https://raw.githubusercontent.com/community-scripts/ProxmoxVED/main/ct/snapotter.sh)"

Getting this error:

/opt/snapotter_data/ai/venv/bin/python3: can't open file '/app/packages/ai/python/install_feature.py': [Errno 2] No such file or directory

Although it might be another error than last time

[MickLesk]

its fixed since 1h

[MickLesk]

#14720 (reply in thread)

[dhoitink]

I did a fresh install this morning. Installing the AI functions is working now! However, when trying to use them, I get an error:

Missing dependency: libGL.so.1: cannot open shared object file: No such file or directory. Install opencv-python-headless, numpy, and Pillow.

[Vakrehus]

This will solve your problem.

apt update
apt install -y libgl1

[dhoitink]

i will give it another snot tomorrow! thx.