Skip to content
A ruby gem to validate image uploading
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Type Name Latest commit message Commit time
Failed to load latest commit information.
.codeclimate.yml Exclude google apis classes from test coverage Feb 26, 2016
.rspec Initial commit Feb 23, 2016
circle.yml Add circle.yml file and remove travis.yml Feb 26, 2016
picguard_logo_rsz.png Add resized logo Mar 14, 2016

Code Climate Test Coverage Circle CI

Picguard guards your application by filtering out the pictures showing inappropriate content. It uses Google Vision API to process the pictures that are being uploaded via your application.

See demo

Why use Picguard?

Imagine a business application that allows user to upload a photo when creating his profile. The user is free to upload any kind of picture (including pictures showing violent scenes or adult content). The most popular solution for that is implementing a feature that allows other users to report the inappropriate content - which means you rely on your users when it comes to the safety of your application - not so great. Another scenario would be: what if for a profile picture an application should only allow the picture showing a human face? The soultion would be often the same as in the first example. Picguard lets you configure your preferences (globally or separately for each model) for image filtering and gives you a clean solution to validate the picture before it's saved.


ImageMagick must be installed and Picguard must have access to it.

If you're on Mac OS X, you'll want to run the following with Homebrew:

brew install imagemagick

If you're on Ubuntu (or any Debian base Linux distribution), you'll want to run the following with apt-get:

sudo apt-get install imagemagick

After the installation you should be able to run which convert in your terminal and you should see for example /usr/local/bin/convert as a response.


Add Picguard to your Gemfile:

gem 'picguard'

And run bundle install within your app's directory.


In your Rails app create a file such as config/initializers/picguard.rb and fill it up with your default configuration. Example:

Picguard.configure do |config|
  config.google_api_key = #{YOUR_GOOGLE_API_KEY}
  config.threshold_adult = "POSSIBLE"
  config.threshold_violence = "LIKELY"
  config.threshold_face = 0.8
Let's go through all of the attributes you need to set up in the config:

google_api_key is the secret key that you use to authenticate to the Google Vision API, you can generate one that is connected to your Google Project. More about generating the key can be found here. If you are completely new to the Google Cloud Platform you should probably start here.
threshold_adult and threshold_violence are the thresholds for the adult and violent content. This is the highest value that you consider acceptable, everything above this level will be categorised as unsafe. For all of the likelihood levels please check this piece of documentation.
threshold_face is the threshold for face recognition. Google responds with a float value from 0 to 1 that reflects how sure the Google API is when it comes to face recognition. Only the pictures with values above your threshold will be categorised as the ones showing human face.


To validate the content of the picture simply add the following validation to your model:

  validates :image, guard: {
    safe_search: true,
    face_detection: true,
    method_name: :img_path

image is the name of the model's attribute that should be validated
guard is the name of the picguard validator
safe_search and face_detection are the flags reflecting what should be validated for given model
method_name value is the name of the image's attribute method that returns image file path

NOTE: Sometimes to return file path it's necessary to chain methods such as avatar.tempfile.path. In such case you need to pass an array of symbols to the method_name attribute (e.g. [:tempfile, :path]). For popular file-uploading gems, Picguard gives you a tool attribute that is used interchangeably with the method_name. More examples below.

Sample validations for popular file-uploading gems


class User < ActiveRecord::Base
  mount_uploader :avatar, AvatarUploader

  validates :avatar, guard: {
    safe_search: true,
    face_detection: true,
    tool: :carrierwave


class User < ActiveRecord::Base
  has_attached_file :avatar,
    styles: { medium: '300x300>', thumb: '100x100>' },
    default_url: "/images/:style/missing.jpg"

  validates_attachment_content_type :avatar, content_type: /\Aimage\/.*\Z/

  validates :avatar, guard: {
    safe_search: false,
    face_detection: true,
    threshold_face: 0.5,
    tool: :paperclip


class User < ActiveRecord::Base
  dragonfly_accessor :avatar

  validates :avatar, presence: true, guard: {
    safe_search: true,
    threshold_adult: 'VERY_LIKELY',
    tool: :dragonfly

Custom usage

We also provide analyze method to analyze images wherever you want:

  image_path: 'path/to/image',
  safe_search: true,
  face_detection: true,
  threshold_adult: 'POSSIBLE',
  threshold_violence: 'UNLIKELY',
  threshold_face: 0.7
  • image_path - provide path to image
  • safe_search - determine whether image should be safe due to adult and violence content [true/false]
  • face_detection - determine whether image should contain human face [true/false]
  • threshold_adult - [optional] provide threshold for adult content different than defined in initializer
  • threshold_violence - [optional] provide threshold for violence content different than defined in initializer
  • threshold_face - [optional] provide threshold for face detection different than defined in initializer

NOTE: the bigger the image, the longer it takes to analyze it. For images bigger than 2000px x 2000px consider using progress bars or background processing solutions like Sidekiq, Resque, DelayedJob, etc.


We use two main dependencies - Google API Client gem to connect with the Goole Cloud Vision API, and the MiniMagick gem to resize bigger pictures before proccessing them via Google Cloud.


The gem is available as open source under the terms of the MIT License.

Core Team

You can’t perform that action at this time.