📷 Sample Node.js Application for the IBM Watson Visual Recognition Service
JavaScript CSS HTML
Latest commit 6140a73 Jan 25, 2017 @kognate kognate committed on GitHub Merge pull request #245 from beth/fix/readme
Add clarity to README, fix broken image link, fix typos


Visual Recognition Demo

Build Status codecov.io

The Visual Recognition Service uses deep learning algorithms to analyze images for scenes, objects, faces, text, and other subjects that can give you insights into your visual content. You can organize image libraries, understand an individual image, and create custom classifiers for specific results that are tailored to your needs.

Give it a try! Click the button below to fork into IBM DevOps Services and deploy your own copy of this application on Bluemix.

Deploy to Bluemix

Getting Started

  1. Create a Bluemix Account:

    Sign up in Bluemix, or use an existing account. If creating an account, you'll have to first create an organization and a space - you should be prompted to do this when you sign in for the first time.

  2. Download and install the Cloud-foundry CLI tool.

  3. Edit the manifest.yml file, and change the - name: visual-recognition-demo line to something unique.

        label: watson_vision_combined
        plan: free
    - name: <application-name>
      path: .
      command: npm start
      memory: 512M
      - visual-recognition-service
    NODE_ENV: production

    The name you use determines your initial application URL, e.g. <application-name>.mybluemix.net.

  4. Connect to Bluemix in the command line tool.

    $ cf login -u <your user ID>
    $ cf api https://api.ng.bluemix.net
  5. Create the Visual Recognition service in Bluemix.

    $ cf create-service watson_vision_combined free visual-recognition-service
  6. Push it live!

    $ cf push
  7. You can now visit <application-name>.mybluemix.net to check out the demo!

See the full API Reference documentation for more details, including code snippets and references.

Running locally

The application uses Node.js and npm so you will have to download and install them as part of the steps below.

  1. Create a file named .env in the root directory of the project with the following content:


    You can see the <api_key> of your application using the cf env command:

    $ cf env <application-name>

    Example output:

      "watson_vision_combined": [{
          "credentials": {
            "url": "<url>",
            "api_key": "<api_key>",
        "label": "visual_recognition",
        "name": "visual-recognition-service",
        "plan": "standard"
  2. Install Node.js. Installing Node JS will also install npm.

  3. Go to the project folder in a terminal and run

    $ npm install
  4. Start the application by running

    $ npm start
  5. Go to http://localhost:3000 in the browser to view the demo!


To view your logs and troubleshoot your Bluemix application, run:

  $ cf logs <application-name> --recent

Environment Variables

  • VISUAL_RECOGNITION_API_KEY : This is the API key for the vision service, used if you don't have one in your bluemix account.
  • PRESERVE_CLASSIFIERS : Set if you don't want classifiers to be deleted after one hour. (optional)
  • GOOGLE_ANALYTICS :Set to your google analytics key, if you want analytics enabled. (optional)
  • PORT : The port the server should run on. (optional, defaults to 3000)
  • OVERRIDE_CLASSIFIER_ID : Set to a classifer ID if you want to always use a custom classifier. This classifier will be used instead of training a new one. (optional)

Changing the Included Images

Sample Images

The sample images are the first 7 images when the site loads. They are called from a Jade mixin found in views/mixins/sampleImages.jade. If you just want to replace those images with different images, you can replace them in public/images/samples and they are numbered 1 - 7 and are jpg formatted.

Custom Classifier Bundles

Adding new/different custom classifer bundles is much more invovled. You can follow the template of the existing bundles found in views/includes/train.jade.

Or, you can train a custom classifier using the api or the form and then use the classifier ID.

Getting the Classifier ID

When you train a custom classifier, the name of the classifier is displayed in the test form.

Classifier ID Tooltip

If you hover your mouse over the classifier name, the classifier ID will be shown in the tooltip. You can also click on the name, and it will toggle between the classifier name and the classifier ID.

You can then use this custom classifier id by placing it after the hash in the request URL. For example, lets say you are running the system locally, so the base URL is http://localhost:3000 and then you train a classifier. This newly trained classifier might have an id like SatelliteImagery_859438478. If you wanted to use this classifier instead of training a new one, you can navigate to http://localhost:3000/train#SatelliteImagery_859438478 and use the training form with your existing classifier.


This sample code is licensed under Apache 2.0. Full license text is available in LICENSE.



Open Source @ IBM

Find more open source projects on the IBM Github Page.

Privacy Notice

This node sample web application includes code to track deployments to Bluemix and other Cloud Foundry platforms. The following information is sent to a Deployment Tracker service on each deployment:

  • Application Name (application_name)
  • Space ID (space_id)
  • Application Version (application_version)
  • Application URIs (application_uris)

This data is collected from the VCAP_APPLICATION environment variable in IBM Bluemix and other Cloud Foundry platforms. This data is used by IBM to track metrics around deployments of sample applications to IBM Bluemix. Only deployments of sample applications that include code to ping the Deployment Tracker service will be tracked.

Disabling Deployment Tracking

Deployment tracking can be disabled by removing require('cf-deployment-tracker-client').track(); from the beginning of the server.js file at the root of this repo.